Aiming at the inaccuracy problem of state-of-charge(SOC) estimation under the complex conditions, an SOC
estimation information fusion framework and a multi-model switch estimation(MMSE) algorithm based on the information
fusion technology are proposed for LiFePO4 high-power batteries. This algorithm carries out the feature extraction and the
pattern classification of the charge and discharge processes, and then re-optimizes the estimation model according to the
different patterns. The MMSE algorithm switches to the matching estimation model to acquire better performance based on
the feature matching. The result of simulation with the running data of pure electric bus shows the feasibility and effectiveness
of the SOC estimation based on the MMSE algorithm.