Flood insurance is an important financial measure for flood risk management. However, a significant protection gap in flood insurance exists in many countries due to high flood insurance costs. Reducing flood insurance costs for both policyholders and insurance companies is crucial for implementing flood insurance. In this study, we derive fundamental mathematical principles for reducing overall insurance costs, including premiums and risk reserves, by introducing the geographic complementarity of flood risk based on portfolio theory. We also propose a reasonable premium allocation framework among multiple individual policyholders using cooperative game theory (CGT), which can benefit all policyholders and achieve stable allocation. The proposed approach is illustrated using China as a case study. The results show that there are low correlation coefficients of flood losses across different provinces and river basins in China, showing high geographic risk complementarity. Compared to independent insurance in each province, national flood insurance can reduce total premiums by 14.8% and total risk reserves by 60.8%. The Yangtze River Basin shows the greatest premium reduction by pooling its internal flood risk, compared with independent insurance by provinces. The Yellow and Huai River Basin shows the greatest premium reduction when pooling national flood risk, compared with pooling risk within individual river basins. In summary, geographic complementarity in flood risk has significant effects on reducing flood insurance costs, implying that China can employ this characteristic to implement a national flood insurance program. Furthermore, the proposed approach can also offer new insights for enhancing catastrophe insurance affordability in more countries.