Abstract: Research in motion of inland water transport ship accurately detection of tracking.Using the infrared video image collection tracking the ship from the ship image,the infrared images of contour shape and texture features are not clear,gray characteristicsused used to describe the goal is not obvious.The traditiona single gray feature detection and tracking method does not effectively describe ship goals of gray characteristics not obvious,causiong the ship detection and tracking accuracy is not high.In order to solve this problem,this paper put forward the fusion feature detection and tracking of ship method.Combine the ship of gray characteristics and movement characteristics fusion to describe movement ship goals,avoiding the traditional single gray characteristics method does not describe effectively characteristics of not obvious ship target.And by using particle filter algorithm construct observation probability model,complete the movement of the ship detection and tracking.Experiments show that features fusion method can describe effectively ship moving targets,accurate detection and tracking of the complete ship,ideal result was obtained.