This paper focuses on a distributed solar energy network composed of multiple clusters, which correspond households equipped with PV units, rechargeable batteries, electrical appliances, and an electric power router. The routers manage the power transfer over the clusters. We obtain the optimal power transfer for the most efficient power utilization using a linear programing model. In the numerical experiments, we use real data of power generation and consumption sampled by monitoring in a local city in Japan. First, the power usage efficiensies are compared among the networks with various consumption distributions over the clusters and various battery capacities using the obtained optimal solutions. As the result, the wider distribution of the consumption amount and an appropriate capacity of battery gives the higher efficiency. Second, the sensitivity analysis on the linear programming model shows when, where and how the optimal solution is affected by the change of generation or consumption amount.