Objective To explore methods for improving the accuracy and efficiency of food dynamic sorting through multi-Delta robot collaboration.Methods Based on a multi-Delta robot food automation production line, a collaborative dynamic sorting method integrating dynamic target tracking, multi-robot task classification, and robot trajectory planning is proposed. By accurately calculating the movement distance of the conveyor belt and combining it with real-time target coordinates acquired by cameras, accurate tracking of the dynamic positions of food items is achieved. A centralized control allocation strategy is adopted to assign tasks scientifically and rationally according to each robot's working status and task priority. Multi-objective comprehensive optimal trajectory planning is realized using an improved third-generation Non-dominated Sorting Genetic Algorithm (NSGA-III) and a fifth-order Non-Uniform Rational B-Spline (NURBS) curve. The performance of the proposed method is comprehensively verified through the construction of an experimental platform.Results The proposed multi-Delta robot collaborative sorting method exhibits excellent performance. In practical operation, it achieves high accuracy, efficiency, and stability in food sorting, with a sorting success rate of 100%, an average sorting time of 0.231 s, an average operational impact of 4.45×10³ (°)/s³, and an average energy consumption of 2.45×10² (°)/s², effectively meeting the requirements for efficient and stable food production.Conclusion By optimizing existing dynamic sorting methods and integrating multiple robots, this approach enables accurate, efficient, and stable food sorting.