In large engineering organizations, multiple projects run in parallel, and their progress is controlled through scheduling. Schedulers often go through the cycles of scheduling: making schedules, executing them, monitoring the status, receiving changes in decisions, and updating schedules again. The schedules easily mismatch with how the projects actually go, and generating a new version takes much efforts. Disagreement between stakeholders sometimes arises. Some faulty decisions are made in a hasty start of projects. This thesis investigates how to boost the efficiency of scheduling. To boost the efficiency of scheduling a group of projects, the factors that constrain the scheduling efficiency need to be investigated. First, various stakeholders are interviewed about the efficiency issues. Second, the main ideas obtained in interviews are tested through a survey. The survey ranked the impact of various factors and measured the time usage on different activities. Third, project tasks that have been finished are analyzed in terms of the cause of delay. The causal relations between issues are built through analyzing the data obtained. The causal relations explain how the causes lead to the problems in project scheduling. The first category of causes lie in the uncertainty in the inputs leads to the uncertainty in the schedules. The uncertain inputs are listed in the following paragraphs. First, the work to be done is uncertain. At the early stages of a project, whether the project will remain selected is uncertain. Not all ideas for projects are able to be supported due to limited budget. Therefore, different project proposals need to be compared in terms of costs and other figures. In the selection procedure, some projects get approved while some are stalled. At the beginning of each project, it is not sure how much the project will cost. The project cost is influenced by the project’s scope, design, and outsourcing strategy, which is not static. The ambiguity in a project’s cost affect whether the project is better than other projects. The client may change the decision in project selection, in scoping and in design. Such changes affect the workload, and the workload affects how long the project work takes. Second, the order in which the projects are processed is seldom stable and clear. Within a large company, the engineering department often receives multiple projects from multiple internal clients. Limited engineering resources lead to clients competing for resource usage. Lack of coordination between clients often leads to changing priority ranking. Even for the same client, there are competitions between new projects and existing projects. Third, the resource supply could also vary. When the demand of resources exceeds the amount of engineers in the organization, the wished deadlines for projects specified by clients become infeasible if internal resources are only used. The engineering party needs to negotiate with clients whether to postpone some projects, or outsource some projects. Clients are usually reluctant to agree to outsourcing because the outsourced work ought to be paid from the fixed budget, according to the financial policy of the company. Such disagreement leads to uncertainty in resource supply. Besides, sometimes emergent tasks appear which occupy the hands of the task performers who were reserved by projects. Fourth, coordination between different tasks is challenging. Engineering projects usually involve cooperation between multiple disciplines, professions, or business functions, since the artifact built through a project consists of various parts delivered through different tasks. Each task is undertaken by different task performers. Different parts of a project’s deliverable have to be compatible, implying the discussion in the interface. Chronologically, the tasks in a project need to be carried out sequentially, and thus one slip in some task may hinder the progress of succeeding tasks. Delays often arise which make the baseline schedule unfeasible. Even when tasks’ deadline is extended, delays remain. Therefore, the actual status of projects varies with changes in such inputs. Dividing such a varying schedule into fine-grained time intervals would generate more uncertain items. The increased amount of uncertain items soon overwhelm the attention of managers. The second group of causes responsible for low efficiency in scheduling is that, schedulers tend to expect the progress on every time interval matches the actual status. Such a pursuit of perfect match leads to unexpected side effects. Each task owner protects his/her commitment date by placing buffers. Later such buffers prove to be used on other tasks. Task performers switch attention between multiple tasks, and postpone the start or progress of tasks with later deadlines. At the last moment for a task, things easily go wrong, leading to delays. Third, the decisions that influence project schedules are not easily made and tend to change: the relevant input information is scattered; misinterpretation of technical design diagram sometimes arise; different stakeholders prefer different design options; etc. Based on the understanding about the causes that constrain scheduling efficiency, measures for coping with such challenges are devised. The solutions include: first, the decisions whose execution depend on low chance preconditions, should be removed from schedules to make the schedule more robust. To prevent optimism bias in setting deadlines, multiple estimation methods should be adopted. Second, some insignificant discrepancies between the baseline schedule and the actual progress, could be tolerated. Tasks had better be coordinated through a relay-race style, letting early completions in some tasks offset delays in other tasks. Third, multi-tasking should be reduced to make schedulers figure out the dedicated progress speed when task performers focus. It is better to promote project progress rather than making task performers busy. Fourth, the updates in information needed in scheduling should be pushed into a shared database as soon as updates arise. The inconsistencies between information should be eliminated. Fifth, rules for resolving disputes should be developed to avoid deadlock or indecision in making some joint decisions. The proposed solution is tested conceptually with the problems aforementioned. The relevant empirical studies which have adopted similar solutions in tackling similar problems are mentioned, serving as evidence for the effectiveness of the solution in this thesis. The implementation plan is discussed, which attempts to provide a step-by-step roadmap towards the changes.