The invention belongs to the field of power selling markets of power systems, and particularly relates to a user maximum demand estimation method based on a comprehensive prediction model, which comprises the following steps: acquiring historical load power consumption data, and establishing a model based on a GDP comprehensive power consumption method, a per capita power consumption method and alinear regression method to predict the annual power consumption of the model; predicting the annual growth rate of the electricity consumption of the user by integrating the prediction models; predicting the maximum load utilization hours of the current month based on historical data and an inter-month correlation method; and estimating the maximum demand of the user in the current month in the current year according to the power consumption of the user in the current month in the last year, the predicted power consumption growth rate and the maximum load utilization hours in the current month. The method is used for providing an estimated value based on historical data when a user declares the maximum demand of the current month, the accuracy of the maximum demand declared by the user can be improved, the total electric charge cost is reduced, and effective reference is provided for dispatching and operation and maintenance of a power grid enterprise.