Vendor managed inventory (VMI) is a popular supply chain system where vendor or supplier take responsibility and decision in managing its customers’ inventory. Two important goals of the VMI are improving service level and maintaining inventory still low and available. Many studies in VMI compare their performance with the traditional system. Unfortunately, studies in improving VMI performance are rare. This work aims to improve VMI by implementing Weighted Round Robin (WRR), a popular scheduling model in computer system, in the replenishment model in VMI. WRR is popular because of its load balancing nature. Environment in this work is two-echelon supply chain. The vendor is a multi-product manufacturer. The customers are retailers. This WRR based replenishment model is then compared with two common replenishment models: (s, S) model and (r, Q) model. In this work, we observe two performance parameters: sales and inventory condition. Based on the simulation result, it is shown that the WRR model performs better than the existing (s, S) model and (r, Q) model and it occurs in most of the observed variables. In the certain condition, performance of the WRR model compared with the (s, S) model and the (r, Q) model is as follows. The WRR model performs 31 percent better than the (s, S) model and 12 percent better than the (r, Q) model in success ratio. Manufacturer’s stock in the WRR model is only 36 percent than in the (s, S) model and 40 percent than in the (r, Q) model. Total stock in the supply chain in the WRR model is only 63 percent than in the (s, S) model and 89 percent than in the (r, Q) model.