Objective: In order to improve the picking efficiency of raw materials in large food processing and storage, a collaborative path planning method of food picking robot based on discrete multi-objective cuckoo algorithm is proposed. Methods: Aiming at the problem of picking multi location raw materials according to the proportion of food processing formula, the cooperative path planning model of multi-objective food picking robot was established, which took the total moving distance, total energy consumption and total weight of periodic picking raw materials as indicators and the Pareto optimal solution discrete multi-objective cuckoo algorithm (DMCA) was proposed. The three-layer cuckoo code was designed, and the renewal strategy of mutation, homogeneous evolution and heterogeneous evolution were redefined to improve the global optimization ability of DMCA. Also, an improved A* algorithm was designed to solve the best moving path between two cargo spaces, and DMCA was used to solve the collaborative path planning model, so that the planned path balanced the moving distance and energy consumption. Results: Compared with other algorithms, the total moving distance was reduced by about 6.3%, and the total energy consumption was reduced by about 7.5%. Conclusion: The proposed method can effectively improve the picking efficiency of raw materials in large-scale grid storage, which has a certain application and popularization value and the planned path balances the moving distance and energy consumption.