This paper adopts the data of recent years of Taipower substation to estimate the parameters of load model by using Measured-based Approach. Using the genetic algorithm finds the optimal parameters. And using fuzzy-genetic algorithm constructs the membership function of all year. Then it can get the optimal parameters of each month for representation the relationship between power and voltage. The results show that using a set of parameters to simulate load's characteristics at each month will get larger mismatch at some month. And simulating by using the parameters of fuzzy-genetic model would get the generally mismatch. It shows the workable of load modeling by using fuzzy-genetic algorithm.