Allseas engineering BV is an offshore company that uses underwater vehicles for conducting subsea operations. The two primary classification of such underwater vehicles are: Remote operated vehicle (ROV) and Autonomous underwater vehicle (AUV). As the names suggests, the former requires a human operator to control and the latter is a fully autonomous vehicle. As of now, the company (and most of the offshore industries) customarily use ROVs. In the recent past, Allseas was inclined to use AUVs as a measure of reducing the operational costs and conducted tests runs with an industrial grade AUV. But the test runs were unsuccessful and the plan of using such an AUV was dropped. It was concluded that the major problem was in launch and recovery operations of the AUV, which were conducted from the deck of a ship. To be specific, the disturbances from the ship’s thrusters, ocean currents, and waves, were proved to be impossible to compensate for by the AUV during the launch and recovery operations. Therefore, this thesis aims to investigate on underwater docking capabilities of AUVs, which not only eliminates the launch and recovery issues but improves AUV’s overall operational capabilities. In this thesis, a docking problem is formulated and the solution to the problem covers the following aspects: modeling of AUV, motion control of AUV, and guidance strategies. For the motion control, an appropriate model of the vehicle is necessary. Various models used in the literature were studied which include: 6 degree of freedom (DOF) non-linear model, 3-DOF horizontal plane model, and 3-DOF vertical plane model. As an example, a 6-DOF non-linear model of ARIES AUV from literature was decoupled into the respective 3-DOF models. These models are used for controller design and docking strategies. A linear parameter varying (LPV) frame work based, gain scheduled feedback and feed forward controllers were developed for the control of vehicle heading and depth. The controller design involves the following steps: Firstly, a third order Quasi- LPV control plant for heading and depth were derived from horizontal and vertical plane models. Then, linear controllers (gains) were designed for fixed values of scheduling variables. Based on the dimension of the scheduling variables, a stability preserving interpolation of these gains were performed to obtain the final controller. A PI controller was designed for controlling the longitudinal velocity of the vehicle. The performance of the controllers were checked on 6 DOF and 3 DOF models of the ARIES AUV. Finally, two docking strategies: Three point and N-point, were developed using lookahead based path following and investigated their performance for two scenarios: stationary dock and non-stationary dock. The docking strategies addressed the docking problem in two aspects: geometric path generation and path following. Finally, the controller performance in closed loop, as in, the guidance, tracking controllers, and vehicle dynamics was compared for the two developed strategies. It was shown that the N-point docking strategy yielded better convergence to the paths than the three point docking.