In order to achieve precise collision avoidance for large ships, a novel intelligent collision avoidance control approach is presented in this paper. To obtain the precise collision avoidance information capability, a fuzzy set interpretation is developed to handle imprecise information. This allows the introduction of an innovative self-training optimizing search method. The optimizing process is based on the particle swarm optimization algorithm and off-line training data that is obtained from trial manoeuvres based on computer simulations. The resulting controller can decrease the ship operators' burden to deal with bridge data and help them to make timely and precise collision avoidance decision. The results show that the designed intelligent controller performed well to implement the optimizing control of ship collision avoidance.