Abstract: The accurate estimation of the state of charge(SOC), state of energy(SOE), state of power(SOP), state of health(SOH) and accurate prediction of the remaining useful life(RUL) of lithium-ion batteries are pivotal to enhancing battery utilization efficiency and safety performance. In order to achieve the aforementioned objective, a multi-time scale estimation framework based on physical data fusion is proposed. The purpose of this framework is to estimate and predict the aforementioned multi-state parameters. Firstly, the first-order RC equivalent circuit model is selected as the physical method basis to achieve accurate estimation of the SOC, SOE and SOP of lithium-ion batteries and explore the coupling relationship between the three. Secondly, according to the voltage, time and incremental capacity(IC) data, the characteristic values are extracted, the SOH estimation model is constructed based on the semi-supervised learning algorithm, and the RUL prediction model is constructed using the long short-term memory network(LSTM) algorithm. Finally, the merits of the equivalent circuit model method and the data-driven method are combined to construct a multi-state joint estimation framework according to the characteristics of the battery state to be estimated. The NASA public battery dataset is then used for experimental verification. The findings demonstrate that the proposed multi-scale estimation framework exhibits commendable performance in terms of estimation accuracy and robustness.