［Purposes ］ To improve the efficiency and service quality of food delivery，a collaborative food delivery service model of drones and riders adopting drone drop -off cabinets was established，and route optimization for food delivery that considered drone and rider collaboration was investigated.［Methods］ To minimize total travel costs of drones and riders，as well as time window penalty costs，a hybrid integer programming model for delivery route optimization was developed，incorporating constraints including drone endurance limits，payload capacity of riders and drones，time windows，pickup -delivery precedence requirements，and flow balance.An adaptive large neighborhood search algorithm was designed，and computational experiments were conducted based on test instances constructed from real -world delivery scenarios，so as to validate the algorithm ’s effectiveness.［Findings］ By comparing different delivery service models，it is found that compared to the traditional rider -only delivery model，the proposed collaborative delivery service model of drones and riders can reduce the average time window penalty cost by 9.10%.As order volume increases，the reduction in the time window penalty cost proportion initially grows but later stabilizes.Sensitivity analysis of drone endurance limits shows that the total costs and time window penalty costs decrease greatly with extended endurance.In addition，sensitivity analysis of payload capacity of riders and drones indicates that increased payload capacity reduces total cost but raises time window penalty cost.Moreover，sensitivity analysis of time window width finds that the total cost and time window penalty cost decrease as the time window gets widened.［Conclusions ］ This study can provide theoretical references for optimizing the food delivery service model of drones and riders and offer route optimization decision -making support for the collaborative food delivery of drones and riders.