The paper proposes an algorithm based on particle swarm optimization (PSO) and intuitionistic fuzzy sets theory. It fuses multiple different neural networks and applies it to the comprehensive assessment of battlefield target damage effect. It adopts PSO algorithm to improve and optimize multiple neural networks of different structure, then confirms the weights of different neural networks, and synthesizes their assessment results as the final output result according to the weight. Apply the algorithm to instance simulation, the result shows its validity and rationality. .