The paper discusses the problems of weapon level year's evaluation , presents a weapon level year's evaluation model based on the back - propagation neural network , and. pre- processes the performance parameters of the weapons concerned ,. The results of simulation show that the model is effective and promising.