Rapid increase of renewable power brings integration challenges, such as need of system flexibility, the spatial mismatch of renewable resources and power loads, and long-term decarbonization policy uncertainty. In this study, a long-term multi-regional power system planning model is proposed with monthly, regional and sectorial demand forecast, describing fluctuation of renewable power, and optimization of installed capacity, power grids and storage facilities. The most cost-effective renewable power deployment of Southwest China is presented as a case study due to abundant renewable resources, significant monthly fluctuation and insufficient infrastructure. Scenario analysis is conducted to address policy uncertainties of carbon tax and power substitution. Results show that albeit limiting and reducing coal power is necessary, de-capacity policies cannot be implemented equally in all regions considering its ability to handle renewable power volatility. To enhance system flexibility, storage facilities are required in Sichuan and Yunnan whilst fossil power plants are required in Chongqing and Guizhou. For cross-region integrating of renewable power, two separate grids in China should be connected between Chongqing and Guizhou. Hydropower could be delivered to eastern regions to meet peak loads of summer cooling. The Southeast power exported to South regions would gradually decrease since its demand increase rapidly. (C) 2020 Elsevier Ltd. All rights reserved.