This paper briefly presents a parametric flight loads identification method which combines genetic algorithm and extreme learning machine (ELM). The model is based on ELM method, and genetic algorithm is used to develop bias weight and weight matrix between input and hidden layer in ELM network. At the end of the paper, GA-ELM model is used to identify flying load based on real flying data. The identify result is compared with that of BP network and original ELM method, and GA-ELM model is proved to be validated, accuracy and feasible.