In the frequency modulation process of a ship diesel-storage hybrid supply system, the energy storage with a faster ramp rate frequently charges and discharges in response to load fluctuations, sacrificing the lifespan of the energy storage while optimizing the frequency dynamic performance. It is a challenging task to consider the differentiation of diesel and storage climbing characteristics and the economic efficiency of energy storage lifespan to achieve the optimal allocation of diesel and energy storage coordinated frequency modulation power in hybrid power management. This paper proposes a diesel-storage coordinated frequency modulation double-layer model predictive control method that takes into account the lifespan of energy storage batteries. Using the active prediction method to track the optimal output of different climbing rates of diesel and storage hybrid power sources, solving the problem that passive control methods, such as AGC, are difficult to simultaneously consider the quality and economic efficiency of frequency modulation. A state space equation of system frequency-diesel-storage is established, and an upper-level power allocation method for diesel-storage coordinated frequency modulation prediction and control is proposed, allocating diesel frequency modulation power according to the response characteristics and operating constraints of diesel and storage. A lifespan economic model of energy storage batteries that considers the state of charge, cycle times, and investment cost is constructed, and a lower-level power allocation method for energy storage system lifespan economic optimization control is proposed, allocating frequency modulation power to each energy storage battery pack based on lifespan economic efficiency. A ship power system simulation model is built using MATLAB/Simulink, and compared with passive control methods such as AGC, the simulation results show that the proposed method simultaneously improves the lifespan economic efficiency of energy storage and the system frequency stability in the dynamic frequency modulation process.