Despite recent progress in clinical study and biological technology of cancer, the prognostic prediction of patients still remains difficult and inaccurate. Since DNA microarrays permits a simultaneous analysis of multiple genes, it has been used to profile gene expression which can categorize cancers into subgroups [1]. To analyze gene expression data, many statistical techniques have been used, but there would be relationships among genes that cannot be expressed statistically. Therefore, artificial neural network (ANN) and fuzzy neural network (FNN) are useful for finding relationships with high accuracy [2]. However, the immensity of data makes us spend much computational time to construct the ANN or FNN model. In this paper, we applied fuzzy neural network (FNN) combined with SWEEP operator method to accelerate the calculation speed about 30 times faster than the FNN modeling method using the back propagation leaning algorithm. The constructed models achieve high accuracy for prognosis of patients. The results in modeling were evaluated to compared with those of Multiple Regression Analysis (MRA).