In day to day life, there are enormous amount of data has been collected, stored, processed and transmitted. The computer revolution and the Internet has brought radical change to an organization's environment. However a group of people access and manipulate information is becoming radically different for the ease of decision making. So, to extract the knowledge from the huge amount of data can be done only by unlocking the hidden data into some useful information. For this we have a best technique called Data mining. Data mining techniques are widely used in classification and prediction in the field of bioinformatics. In a medical diagnostic system where we wish to predict a patient disease based on various laboratory test, where some of the lab test results may not be essential to diagnosis the problem. This created a seed for us to go for the task of feature selection using Chi-Square attribute evaluation which can deal with the identification and elimination of noisy features and to provide accurate and quality diagnosis to the patients and also reduced the amount of storage in the databases. © Research India Publications.