Energy storage is the core technology for building new power systems, among which, electrochemical energy storage of lithium-ion batteries is the main form at present and is of great significance to achieve the goal of "double carbon". Fault diagnosis is of great sig-nificance to ensure the safe operation of battery energy storage system, especially the accurate diagnosis of minor faults can effectively prevent the occurrence of serious faults, however, the traditional fault diagnosis method has poor timeliness and low accuracy, and it is difficult to capture the characteristics of minor faults. Therefore, a probabilistic neural network (PNN) improved by applying the sparrow search algorithm (SSA) was proposed for the diagnosis of minor faults in energy storage batteries. First, the fault characteristics were analyzed by the lithium-ion battery fault categories to extract the state feature information after the occurrence of minor faults. Then, the lithium iron phosphate energy storage battery fault signal was decomposed into a series of feature vectors and input into the SSA-PNN model. Finally, an experimental validation study was carried out. The results show that, compared with the traditional fault diagnosis method based on error back propagation algorithm, the SSA-PNN-based fault diagnosis method achieves 99.7% accuracy with higher diagnostic accuracy and real-time performance.