Abstract: Gravity energy storage is a novel physical energy storage technology. To address the problems including discreteness, fluctuation and insufficient power rating in the power characteristics of single-unit gravity energy storage operation, this paper focuses on the key technologies for the coordinated control of hybrid energy storage, and designs a gravity-flywheel-battery hybrid energy storage system as well as its grid-connected control strategy. Firstly, a motion model for gravity energy storage is established based on the system operation principle, and the mathematical model of the permanent magnet synchronous motor in the dq coordinate system and the time-sharing startup control strategy for multi-unit gravity energy storage are constructed. Secondly, a speed/power-loop control strategy is designed for the machine-side converter to regulate the motor speed and torque so as to simulate the operating process of gravity energy storage. A voltage-loop control strategy is developed for the grid-side converter to maintain the stability of the DC voltage. To cope with the inherent power fluctuation of gravity energy storage, flywheel and battery are adopted to perform power compensation for gravity energy storage, and the power-loop converter control strategies are designed for flywheel and battery. Finally, the grid-connected model of the hybrid energy storage system is built on the MATLAB/Simulink platform to verify the feasibility of the constructed models and control strategies. Simulation results show that the proposed converter control strategies are effective; and the established hybrid energy storage system achieves stable grid-connection and can further smooth grid-connected power fluctuations. The flywheel energy storage, battery energy storage, and hybrid energy storage can respectively reduce the grid-connected power fluctuation rate to 4.00%, 4.73%, and 2.95%, while the levelized cost of electricity reaches 0.810 ¥/(kW·h), 0.785 ¥/(kW·h) and 0.787 ¥/(kW·h). The hybrid energy storage scheme takes into account both response speed and economic performance, and presents the optimal performance for engineering applications.