Car-sharing is a travel mode that can serve as an alternative to private cars, helping to reduce urban pollution. However, currently, there is a low willingness among travelers to use car-sharing, which is reflected in both low market penetration and user frequency. Therefore, it is essential for the government to encourage the use of car-sharing by providing subsidies. To better encourage the usage of car-sharing, this paper applies a two-fold evolutionary game model involving travelers and the government to explore the impact of subsidies on travelers’ choices, and the factors that could affect the subsidies’ efficiency. A simulation, using data from Beijing, was conducted to determine the implications of subsidy policies. The results show that a mileage-based subsidy and a fixed subsidy are applicable to travel of high and low mileages respectively, and under both subsidy modes, subsidies for trips with short duration or short pick-up and return time are more effective. Furthermore, we find that the efficiency of subsidies increases as the scale of car-sharing users, demand elasticity, or total number of travelers increases. Additionally, the subsidy levels should be lower than the environmental benefits of car-sharing but higher than the difference in travel costs between private cars and car-sharing. Future work will involve other game players such as car-sharing operators in order to draw deeper conclusions, and will involve the collection of data from more countries and cities to develop the robustness of the conclusions.