In the light of the characteristics of the data on power load， a database is constructed and the association rules data mining method is applied to the urban power load analysis. The FP-Growth algorithm corresponding to the FP-Tree structure（frequent pattern tree） is then used to analyze the influence of correlative factors on power load through an analysis of the load data on 48 cities in China. The results reveal the relations between annual power consumption and correlative factors such as the GDP growth rate， the proportion of the second industry， the center hierarchy， the executive category of a city， etc. The results agree with practice while giving a specific numerical interval of factors.