Abstract: Aiming at the problems of low quality, complex operating conditions, lack of complete charging and discharging cycle data and multi-source heterogeneity of the energy storage battery operation data collected by the existing battery management system (BMS), we studied the mechanism of splicing and reconstruction for lithium-ion energy storage battery operation data by using the gradient descent algorithm, and proposed a method of data splicing and reconstruction suitable for the characteristics of lithium-ion energy storage battery operation conditions. Based on the analysis for the transient and steady state characteristics of the battery when the operating conditions change, the mechanism of splicing and reconstruction for the lithium-ion energy storage battery operation data are revealed. The empirical mathematical equations are established by combining the main operating parameters such as voltage, current and capacity. The empirical values and boundary conditions related to the mechanism of battery operation data splicing and reconstruction are clarified, and the evolution law of operating conditions and current rate is studied. On this basis, a new method of splicing and reconstruction for the lithium-ion energy storage battery operation data is proposed. The experimental data and real operation data verify that the method has high accuracy and adaptability.