The rise of the sharing economy has promoted the rapid development of the shared mobility industry. Carpooling has become a new trend in mobility. It is helpful to consider the service mode preference of drivers and riders in carpooling to meet diversified consumer demand. A new matching method is proposed to address the driver-rider matching problem for carpooling with service mode preference. First, the driver-rider matching problem for carpooling with service mode preference is described. Then, the satisfaction of riders and drivers is calculated based on the time window and maximum detour distance. Furthermore, a multi-objective optimization model is constructed with the objectives of maximizing the number of matches, maximizing the driver’s average satisfaction and maximizing the rider’s average satisfaction. Aiming at the NP-hard characteristic of the model, a heuristic algorithm for solving the model is designed based on the non-dominated sorting genetic algorithm II. Finally, we conduct numerical experiments to validate feasibility and effectiveness of the proposed method. The results show that the method can effectively solve the driver-rider matching problem for carpooling with service mode preference, and the initial population generation strategy and local search operation can enhance the global search capability of the algorithm to some extent.