To address the problems of over-maintenance and delayed maintenance in traditional maintenance management models following the deployment of large-scale high-tech equipment in military forces， which make it difficult to detect equipment faults in a timely manner and seriously affect the generation of equipment combat effectiveness and the accomplishment of military missions， this paper proposes an improved grey GM（1，1） model based on an HHO-PSO hybrid optimization algorithm that integrates Harris Hawks optimization and particle swarm optimization. By synergizing the global exploration capability of HHO and the local feature extraction capability of PSO， the algorithm optimizes parameters and establishes model adaptation using root mean square error （RMSE） as the convergence criterion， thus effectively improving prediction accuracy. The experimental results indicate that the improved model demonstrates good performance in equipment fault prediction， with low errors and high accuracy， and can provide technical support for early fault warning of equipment.