Firstly, a mathematical model of main engine cooling water temperature regulation system is established according to the working principle of the high temperature fresh water cooling system. Secondly, the aforementioned system is chosen as the research object and a neuron-based adaptive control system is applied to regulate the temperature of cooling water. According to the algorithm of the control system, the neuron makes self-organizing through modification of its weighted values, and the weighted values are determined through correlative search. Additionally, a smoothing factor is applied during the modification of the weighted values, which makes the procedure of the weighted value modification more smoothly. The experimental results show that the dynamic property of the cooling water temperature regulation process is good and the neuron-based adaptive control system is suitable for control of main engine cooling water temperature.