An adaptive single user flexible load control evaluation algorithm is proposed to design an evaluation scheme adapted to various demand response scenarios and improve users’willingness to participate in regulation and control. Firstly, the identification method of target load curve is proposed, and then the generalized distance between curves according to the characteristics of common evaluation indexes is defined. By adjusting the reserved parameters in the generalized distance, the evaluation scheme can be adapted to different scenes. Then, the method of setting initial parameter values is proposed. Using this method and the neural network, combining with the existing evaluation samples,each parameter in the generalized distance can be confirmed adaptively.Lastly, the effectiveness of this algorithm is verified based on the data of Lvjian area in Changzhou.