An improved entropy weight algorithm is introduced into a grey relational analysis algorithm to overcome the shortcomings of similar influencing factors, and the reasonable weights of each influencing factor are obtained to improve the objectivity of the evaluation. Using the improved entropy weight-grey correlation analysis algorithm, the correlation values between the reliability index of the power supply area and each influencing factor are obtained. In the experiment, as for the data of four power supply areas, the main factors affecting the reliability of power supply in different power supply areas are obtained by using the method proposed in this paper. The results are compared with those obtained by the gray correlation analysis algorithm. The experimental results show that the method can obtain the influences more effectively. Thus, the key factors that have a significant impact on the reliability of the power supply in a power supply area are determined. This work is supported by National Natural Science Foundation of China (No. 61503237).