In order to realize the dynamic assessment of battery state under the whole life cycle of energy storage batteries, and to improve the adaptability of the lithium-ion battery model and the accuracy of state estimation under complex working conditions, a joint estimation method of battery state of charge (SOC) and state of health (SOH) based on the improved technique for order preference by similarity to an ideal solution (TOPSIS)-fuzzy Bayesian network is proposed. The equivalent circuit model of the battery is constructed by applying the multi-order resistor-capacitance circuit (RC) model and the node-branching framework, and the parallel loop in the equivalent circuit model of the second-order RC battery is characterized by Kirchhoff's law and Ohm's law to construct the spatial equations of state and the equivalent output equations. The constructed equations of state are discretized, and the discretized state-space equation of the battery model is analyzed by defining the discretized zero-input response and zero-state response of the parallel independent loop. The expert scoring method is introduced into the TOPSIS algorithm for the quantitative estimation of battery SOC, and combined with the Bayesian network that integrates into the fuzzy scale, the corresponding SOC values in the observed samples of the batteries are calculated from the battery SOH values under the same time distribution scale, so as to realize the joint estimation of battery SOH and SOC. The experimental results show that the proposed method can effectively estimate the results of battery SOC and SOH in different discrete spatial scales, and the estimation method has good accuracy and high precision.