Classifiers are often data dependent as they perform better on one type of data, but fail to perform well for another data set. There is a need for robust classification algorithms which exhibit performance stability for multiple types of data. This problem can be addressed if different classifiers are fused to identify a particular class. In this paper, we have implemented the idea of classifier fusion using six different classifiers to classify the microarray gene expression data of breast cancer patients. The paper uses two classifier fusion models: majority voting and random bagging to improve the accuracy of the classifiers. Our experimental results have shown that the new proposed classifiers fusion methodology have outperforms single classification models.