Due to different application scenarios and diverse load devices, a single factor such as accuracy rate is likely to cause the evaluation results to be deformed and not easily detected, so that the performance of the non-intrusive algorithm cannot be comprehensively and effectively evaluated. Therefore, this paper proposes a TOPSIS model based on fusion decision-making. The evaluation system is constructed with multiple evaluation indicators. The fusion weights based on AHP and coefficient of variation are adopted, taking expert experience, engineering requirements and objective data laws into account. The method avoids a single subjective assignment that causes significant differences in indicator data to be ignored and a single objective assignment to cause data error fluctuations to be exaggerated. Finally, the ranking method of approximating the ideal solution is utilized to calculate the closeness of the evaluation object with the positive and negative ideal solutions, and the algorithm comprehensive sorting result is obtained. The experimental results show that the evaluation model proposed in this paper can effectively evaluate the performance of the algorithm and provide a new solution for the comprehensive evaluation of non-intrusive load identification algorithms.