Objective: In order to solve the problems of low efficiency and poor precision of parallel robot in food sorting. Methods: Based on the structure of the food sorting system, a moving target grasping strategy of delta robot based on Improved BP neural network and PID control is proposed. The improved particle swarm optimization algorithm is used to optimize the initial weight of BP neural network, and the optimized BP neural network is used to adjust the PID control parameters in real time. The performance of this method is analyzed by experiments, and its superiority is verified. Results: Compared with traditional control methods, the proposed method can achieve dynamic target capture more accurately and efficiently, and the success rate of dynamic capture is more than 98%, which can meet the needs of food sorting. Conclusion: The grasping efficiency and accuracy of delta robot can be effectively improved by optimizing the grasping strategy of moving target.