We study the renewable energy generations in Hong Kong based on realistic meteorological data, and find that different renewable sources exhibit diverse time-varying and location-dependent profiles. To efficiently explore and utilize the diverse renewable energy generations, we propose a theoretical framework for the cooperative planning of renewable generations in a system of interconnected microgrids. The cooperative framework considers the self-interested behaviors of microgrids, and incorporates both their long-term investment costs and short-term operational costs over the planning horizon. Specifically, interconnected microgrids jointly decide where and how much to deploy renewable energy generations, and how to split the associated investment cost. We show that the cooperative framework minimizes the overall system cost. We also design a fair cost sharing method based on Nash bargaining to incentivize cooperative planning, such that all microgrids will benefit from cooperative planning. Using realistic data obtained from the Hong Kong observatory, we validate the cooperative planning framework and demonstrate that all microgrids benefit through the cooperation, and the overall system cost is reduced by 35.9% compared with the noncooperative planning benchmark.