Abstract: In this paper, firstly, a sample set security state data is established based on the big data analysis to test the dimensions of the security state of the electric power theft detection and prevention system, and the security coefficient is obtained after resting the system.Then, according to the correlation factors between the electric power theft detection and prevention system and the data set, the neural network algorithm is utilized to input the security data in the input layer, and the output detection results are obtained after processing calculations in the hidden layer.Finally, the technical loss of power theft is calculated and the measured ohmic loss is obtained, and the design of the power theft detection and prevention system is completed based on the iterative training by analyzing the big data.The results show that the trend of electricity consumption by normal users and electricity theft reveals that the electricity consumption of electricity theft can reach up to 1.5 kWh after the electricity is used for about 12 days, which is significantly higher than the electricity consumption of normal users.Therefore, the power theft detection and prevention system based on big data analysis can effectively detect power theft based on user data with high accuracy.