The stochastic demand inventory routing problem (SDIRP) is a typical NP-hard problem. It is also the key to implementing vendor managed inventory (VMI) strategy, that is, to coordinate the inventory problem and distribution problem in a stochastic demand environment. This paper studies the SDIRP based on the Fixed Partition Policy (FPP). Under this policy, customers are partitioned according to the service regions they are in, and customers who are in the same service region are served simultaneously. In this paper, a modified C-W saving algorithm is designed to partition customers, taking into account the impact of partition policy on inventory costs and distribution costs. It is shown that the optimal inventory policy for individual service region is a (s,S) policy, whereas the inventory policy for customers in each service region is an order-up-to policy. Furthermore, this paper proposes an algorithm to solve SDIRP based on FPP. Finally, a numerical example is presented to confirm the efficiency and applicability of the proposed algorithm.