Nowadays, electrochemical battery storage systems are so important in both stationary and mobile applications, especially for telecommunication fields. The lead-acid batteries hold the highest share in installed batteries of all storage technologies because of their maturity, simplicity of manufacturing, and their low cost, which has always attracted telecom industries. However, this technology has limitations, such as poor energy density, low performance, hard maintenance, and short lifespan. Recently, both Lithium NMC (LiNiMnCoO2) and Lithium LFP (LiFePO4) are introduced as new energy storage technologies that promise long-lasting, better efficiency, and higher energy density compared to the old generation of storage systems. Telecom industries are looking for an optimal combination of different criteria such as operating conditions, safety, cost, and effectiveness. In this context, this technical paper presents firstly a mathematical data-driven model to estimate the round-trip efficiency for a battery module for real time predictive control and optimization purposes. Then, this model has been validated using various experimental tests and measurements for different battery module technologies used in telecommunication radio access networks (RANs) under several loading conditions. The testing procedure and results are also discussed.