In this study, a neural network based approach is used to predict the presence of Maturity Onset Diabetes type 3, referred as MODY3 Type II diabetes mellitus. The study has used backpercolation neural network algorithm to predict the specific genetic mutation that causes the MODY3 type II diabetes mellitus. A set of coded numeric values are assigned for numeric representation of genetic data that are available in public domain repositories. A point mutation is introduced in a portion of the nucleotide for the mutation prediction to train the data set. The study has demonstrated that backpercolation neural network algorithm is useful to train and to predict gene point mutation that leads to MODY3 type II diabetes.