Maritime surveillance system installed on-board is crucial in protecting commercial vessels worldwide, which suffer potential losses due to the maritime piracy assault. On-board camera provides a way of monitoring suspicious activities at a low cost. However, the maritime environment poses challenges in detecting mobile targets because the water background is highly dynamic. This thesis addresses the problem of detecting and tracking mobile targets in the maritime environment utilizing an on-board camera. An approach based on optical flow is used to detect mobile targets presented in the camera scene. The camera measurement model and target kinematics model are proposed such that targets can be tracked utilizing the Bayesian filtering technique. A sensor planning strategy based on expected entropy reduction (EER) is developed to select optimal sensor field of view (FoV) locations such that the expected entropy reduction is maximized.