This article investigates the design and optimization of heliostat fields in solar thermal power generation technology. We proposed an optimization method based on the PSO-BP neural network model and analyzed and validated it. By introducing the principle and key parameters of the heliostat field, the main optimization problem was proposed: redesign the parameters. To address these optimization issues, we proposed an optimization method based on the PSO-BP neural network model. PSO (Particle Swarm Optimization) is a heuristic optimization algorithm, while BP (Backpropagation) is a training algorithm for neural networks. By combining these two methods, we can effectively optimize complex problems. We first defined a neural network model, where the input layer includes parameters of the solar collector, and the output layer is the annual average output thermal power of the unit mirror surface area. Then, we used the PSO algorithm to optimize the weights and biases of the neural network to achieve the optimal output. To validate this method, we conducted simulations and tests using actual data. The results show that this method can effectively improve the performance of model, enhance the efficiency and economy of the system. This work provides valuable reference for the development of solar thermal power generation technology.