【摘要】The online-to-offline (O2O) community supermarket is currently a popular O2O business model in China. Owing to the small lot-size, high frequency, time-sensitive, and dynamic arrival of online customer orders, many O2O community supermarkets face challenges in how to pick up the dynamic arrival orders and deliver them to customers with minimum makespan and delivery cost. To achieve the global optimal order fulfillment performance, we study the online integrated order picking and delivery problem for an O2O community supermarket, and order pickers’ learning effects are considered to better plan the integrated problem. To propose a feasible and efficient schedule, the online algorithm A is established, and the competitive ratio is proved to be 2 theoretically. To further verify the effectiveness and efficiency of algorithm A in practice, we summarize the actual order fulfillment rules (named A1), and conduct numerical experiments to compare algorithm A with A1. Moreover order pickers’ workforce characteristics are varied to evaluate the learning effects on the order fulfillment process. The results show that Algorithm A performs better than A1 in different situations, and considering pickers’ learning effects is significant for the accuracy and predictability of order fulfillment process.