In smart grid, most of the current demand response (DR) management methods do not consider user constraints when reducing peak load through data analysis. To solve this problem, a data analysis demand response (DADR) management scheme for residential load is proposed to reduce peak load demand. The proposed method is based on the analysis of power consumption data collected from smart homes (SHs), taking into account the factors such as electrical regulation factor, electrical priority index, electrical power curtailment priority, etc. Based on these factors, different algorithms are proposed to make DR decisions in peak load scenarios. The proposed scheme is tested with open data set, and the test results verify the effectiveness of the proposed scheme. The proposed method can effectively reduce the peak load of power grid, make it meet the power generation capacity limit, and reduce the total electricity cost of users.