The invention provides a data processing method for multistage pull-down question types in an online questionnaire survey system. The data processing method comprises the following steps: S1, inputting questionnaire data; S2, processing questionnaire data; S3, reading format data stored in the step S2, converting the format data into a user interface, taking the user interface as a questionnaire mode, and performing multi-level pull-down selection; S4, inputting answer sheet data; S5, processing questionnaire data; S6, reading all answer sheet data in S5, grouping the data by using group, grouping the data according to multi-level pull-down answers, and counting the frequency of each answer; S7, splitting the multi-level pull-down answers by using spit, carrying frequency, and converting the multi-level pull-down answers into a list array; and S8, outputting. According to the scheme, storage is simpler, the query is quicker, and the performance of the scheme can be better manifested along with the increase of the number of data sources, the increase of answer sheets and the increase of hierarchies.