Abstract: As the recorded data of load characteristic recording devices increase, the calculation work of load modeling using genetic optimizing algorithm increases dramatically. Referring to synthesis load modeling method, the idea of integrating composite load models in different time interval is proposed to take full advantages of the previous identification results and to reduce identification process. Detailed recurrence formulas are given. The weight coefficient selection principle is discussed. Owing that the new data can reflect load characteristics better, weight coefficient selection principle according to recorded time of data are given. Pratical load modeling shows that this method is simple, effective, time saving and can reflect the time variant load.