This paper addresses the problem of adaptive multi-sensor fusion. Firstly, in order to solve this problem, the model of the marine craft and multi-sensor are built. Then the desired path is known, at the same time the north east coordinate is found. In order to get the information of the marine craft, the DGPS is used. Then the information about ship, such as the position, direction and speed, is attained. According to this information, the fusion algorithm will be found. In order to make the algorithm adaptive to the noise system, the noise estimator is used to statistic system noise. This paper proposes CMACKF distributed algorithm to solve this problem. Furthermore, the simulation is used to verify the effectiveness of the algorithm. The simulation results demonstrate that the performance of the proposed algorithm is significantly superior to that of the standard CKF. That leads to improve navigation precision.