The invention belongs to the technical field of geoscience big data deep mining methods, and discloses a clustering and association rule analysis-based geoscience big data deep mining method, which comprises the following steps of descriptive analysis, diagnostic analysis, predictive analysis and instructive analysis, a large amount of data information obtained through investigation is preliminarily sorted and concluded so as to find internal rules, namely a centralized trend and a dispersed trend, of the information, and single-factor analysis is carried out mainly by means of statistics such as a mean value and a percentage represented by various data. According to the geoscience big data deep mining method based on clustering and association rule analysis, a lot of intelligent, deep and valuable information is obtained through analysis of geoscience big data, new geoscience data or new geoscience knowledge and knowledge discovery are generated through analysis, comparison or processing, unknown and hidden data are discovered, and the knowledge discovery is also realized. And the sublimation from'data 'to'knowledge' is realized.