In order to increase the secure storage capacity of university financial system process operation data under the blockchain environment, a secure cloud storage algorithm of university financial data based on the blockchain technology is proposed. The blockchain storage structure model of university financial system process operation data is first constructed, and then the mapping method of Atlas features is adopted. Finally, the blockchain equilibrium configuration parameter analysis model of university financial system process operation data is established. According to the outcomes of the feature extraction process of cloud resource storage Atlas of university financial system process operation data, the fuzzy clustering method is implemented to comprehend the rational planning of cloud storage space. The resource cloud storage structure configuration of university financial system process operation data in cloud environment is realized by using the method of block link fusion and channel equilibrium configuration. The multi-layer modal structure decomposition and fuzzy clustering processing are carried out on university financial system process operation data storage information in blockchain environment by using the empirical mode decomposition method. According to the data graph clustering results, the cloud resource graph of university financial system process operation data becomes smooth in the adjacent wave domain through cloud information fusion and block clustering, which effectively reduces the data storage overhead and improves the secure cloud storage capability of university financial data. The simulation outcomes indicate that this approach can significantly increase the storage performance of process operation data of university financial system under blockchain environment, with better data classification storage, internal structure information fusion performance of university financial data, and lower storage overhead than other methods. We observed that this improvement, in terms of storage overhead costs, can be as high as 43.67% higher than the wavelet method and 30.45% higher than the mode decomposition approach.