Existing analytical and simulation models for lithium-ion battery thermal behavior encounter efficiency or accuracy challenges in energy storage applications. This paper proposes a thermal analysis method for pouch lithium batteries considering dynamic and non-uniform thermal characteristics, providing an online state evaluation tool for energy storage systems. First, a thermal circuit model and analytical partial differential equations are created to capture the dynamic and non-uniform thermal characteristics during charge and discharge. Second, an iteration mechanism is formulated to determine the solving parameters, and a solution algorithm with balancing accuracy and efficiency is proposed. The proposed model and algorithm are applied to the temperature evaluation of commercial energy storage lithium batteries. Experimental results under various charging/discharging conditions show that the proposed method has a millisecond computational cost and less than 3% temperature error. The parameter calibration and iteration mechanisms enhance the engineering applicability. Thus, it demonstrates good application potential for the battery management system of energy storage.