Abstract: Reasonable order batching and fast picking of goods are of great significance to the timeliness of distribution.To study the order batching problem in the O2O supermarket warehouse distribution center so as to allow the picking staff with labor heterogeneity to pick the ordered goods faster and ensure their high efficiency and appropriate amount of workload，a dual-objective mixedinteger planning model is built，aiming to minimize the walking distance of the picking staff and optimally balance their workload.A dynamic adaptive step-size firefly algorithm is applied to address the problem，and the correlation between the proposed model and the objective is proved.Finally，the validity of the model and algorithm is verified through data experiments，showing the model exerts a strong influence on order batching and employees with labor heterogeneity and provides a scientific basis for improving the timeliness of supermarket delivery.