Biclustering gene expression analysis problems are important research problems for applications of DNA chip. They can be applied to new drugs discovery, gene expression data analysis and gene sequencing analysis. Current approaches for biclustering problems can not handle huge amount of gene data and are difficult to be used to solve all types of biclustering problems. In this research, we propose the use of frequent itemset mining to solve the biclustering problems. Our approach has the advantage of easy implementation, low implementation cost, and high integration capability. Experiments on breast cancer data show that our approach can identify more quality biclusters than the existing approach.