The invention relates to a mass historical data mining method of a power distribution network dispatching control system. The mass historical data mining method is technically characterized by comprising the following steps: step 1, performing discretization processing on historical data; step 2, performing data mining on the historical data of the cross section; step 3, mining out a frequent itemset of the distribution network database; step 4, mining strong association rules of various types of historical data; step 5, checking whether the strong association rule has strong correlation or not based on a lifting degree algorithm; step 6, establishing various strong association rules in the historical data of the power distribution network; step 7, judging whether the established strong association rule is reasonable or not; step 8, checking the historical section database by using a strong association rule; step 9, calculating an evaluation result of the distribution automation system; and step 10, after data mining is completed, quitting the execution logic. According to the method, the influence of inconsistent data records in a historical database on distribution network operation evaluation can be reduced, and the distribution automation level is improved.