Multi-stream receive-spatial modulation (MSR-SM) technology is applied to the cell-free multiple-input multiple-output (MIMO) downlink transmission, where both users and access points (APs) are multi-antenna, providing a new idea for green and energy-saving wireless communication. The power control coefficients of each AP are obtained using the sequential convex approximation (SCA) algorithm to optimize energy efficiency (EE) under the constraints of the highest transmitting power of the AP and the lowest signal interference-to-noise ratio of the user. This paper proposes an access point selection scheme based on a water-filling algorithm to reduce the number of unnecessary service APs and the total power consumption of the system. The simulation results show that the proposed power allocation algorithm outperforms the max-min power control algorithm in terms of energy efficiency. The MSR-SM-based downlink transmission mode is more energy efficient than the traditional space division multiplexing MIMO digital modulation mode. Furthermore, the SCA algorithm based on AP selection can improve energy efficiency performance compared with the power control algorithm without AP selection.