Aiming at the problem of how to improve the service satisfaction quantity of the self-powered roadside unit, a transmission control strategy to maximize the number of satisfied vehicle service requests under the constraint of average energy efficiency is proposed. The strategy makes the transmission scheduling decision with the vehicle position and the energy queue length of the self-powered roadside unit as the system state according to the weight order of the vehicles to be served. By establishing a Markov chain model for self-powered roadside unit transmission scheduling decisions, the average number of requests completed and the average energy efficiency in the service process are analyzed, and a nonlinear optimization problem is proposed and solved to reveal the optimal transmission control strategy and its scheduling parameters. The simulation results show that the obtained transmission control strategy of self-powered roadside unit has a double threshold structure with position state and energy state. It has the advantages of energy efficiency and stability. Compared with the greedy strategy and Q-learning algorithm, the energy efficiency is increased by 20.55% and 11.86% respectively. In terms of service stability, compared with the other two strategies, the outage probability and non service probability are reduced by 20.03% and 15.14% respectively.