It is one of the important issues for electric vehicle to utilize power batteries which have long lifetime and excellent performance. For optimizing electrochemical performance and lifetime of the lithium ion battery, an electrochemical-thermal model based on dynamic response is developed by COMSOL MULTIPHYSICS. The modeling theory is the reaction mechanism of lithium iron phosphate battery which also includes a parasitic reaction occurring in the constant current and constant voltage charging process. The model consists of three parts: electro-chemical model, thermal model and capacity fade model. A series of temperature-dependent parameters and lithium ion concentration-dependent parameters relevant to the reaction rate and Li+ transport are employed in this model. Comparing with the results of the experimental test, the model shows high accuracy and reliability. The capacity losses and electrochemical behaviors of the battery in cyclic processes with different rates are investigated. The results show that when the battery is cycled at a rate of 1C, the capacity fading rate is about 6.35%, meanwhile the solid electrolyte interface membrane impedance of the battery is increased by 15.6 mm-2 after 800 time cycle. In the charge process, the side reaction rate within the anode shows a decreasing trend along the direction from the cooper current collector to separator, which is consistent with the lithium concentration in the anode. Besides, the effects of charge/discharge rate, negative active material particle radius and negative solid volume fraction on the battery cycle life are also discussed respectively. Compared with the fading rate of 3.31% after 400 time cycle with 1C rate, the capacity fading rates for 2C, 3C, 4C reach to 3.93%, 4.69% and 5.04% respectively. When the average particle radii of the anode are 2 m and 10 m, corresponding capacity fading rates are 2.89% and 3.87%, showing a difference of nearly 1%. The study for solid volume fraction demonstrates that the battery with a solid volume fraction varying in a range of [0.5, 0.6] will keep a longest battery life. These results show that the model has great potential to optimize the design of the battery.