Abstract: To address the multi-product inventory routing problem with split deliveries (MIRPSD), a vehicle routing optimization model is established for minimizing inventory holding cost, transportation cost and vehicle usage cost. The delivery schedule and the transported quantity of each product for each customer are also considered in the model. A hybrid genetic algorithm (HGA) is proposed to solve this problem, incorporating a perturbation strategy to improve the search efficiency. Experiments are conducted to select appropriate parameters. Furthermore, the ratio of average daily demand to vehicle capacity and unit inventory holding cost are analyzed to investigate their impact on the split delivery strategy and total delivery cost. Multiple test instances demonstrate that the proposed model and algorithm can effectively solve the problem. When the demand follows a normal distribution and the ratio of average daily demand to vehicle capacity is 0.55, the split delivery strategy can reach the best result. The study extends the theoretical framework of the inventory routing problem (IRP), providing a new perspective for solving MIRPSD and offering theoretical support for relevant decisions of logistics enterprises.