Association mining is a well explored topic applied to various fields. In this article, the associations among the genes have been identified from microarray gene expression data. Here a methodology, called Fuzzy Correlated Association Mining (FCAM), is developed for identifying the associations among the genes that have altered quite significantly from normal state to diseased state with respect to their expression patterns. This idea leads to predict the disease mediating genes along with their altered associations. The proposed methodology involves generation of fuzzy gene sets, construction of fuzzy items, computation of fuzzy support for fuzzy items and fuzzy correlation coefficient of a pair of fuzzy items, generation of associations, and identification of altered associations from normal to diseased state. The concept of finding fuzzy correlation between two groups of items, generation of altered associations among the items (groups of items) and then rank these items (groups of items) according to their importance are the novel contribution of the present article. The effectiveness of the methodology has been demonstrated on five gene expression data sets dealing with human lung cancer, colon cancer, sarcoma, breast cancer and leukemia. As a result, some possible genes, like IGFBP3, ERBB2, TP53, HBB, KRAS, PTEN, CALCA, CDKN2A, has been found as important genes that may mediate the development of various cancers considered here. For comparison, we have considered 11 existing association rule mining algorithms. The results are appropriately validated in terms of gene–gene interactions, functional enrichment, biochemical pathways, and using NCBI database.