The invention provides a battery state prediction method and system. The method comprises steps of: establishing a random data assimilation function model I*t = f(U, I, T) representing a function relationship between the dischargeable time t of a battery and voltage U, current I, and temperature T; computing the complete discharge curve Ik*t = f(U, Ik, Tk) and the dischargeable time tk:tk= f(Uk, Ik, Tk)/Ik of the battery to be tested on a current operating condition by using the random data assimilation function model I*t = f(U, I, T) and according to the voltage Uk, the current Ik, and the temperature Tk, measured in real time, of the battery to be tested; and predicting the SoC value and the SoH value of the battery to be tested according to the complete discharge curve and the dischargeable time of the battery to be tested on the current operating condition. The battery state prediction method and system may perform assimilation treatment on battery parameters on different operating conditions, adapt to operating conditions with continuously changed discharge current and ambient temperature, may accurately predict the state of health and the state of charge of the battery, and is good in instantaneity, high in accuracy, and easy to implement.