The need for high speed, low cost and smaller area has increased the integration of electronic devices. As the number of components increases, the reliability of system becomes a major challenge. The bipolar junction transistor is an immensely used passive component in the various electronics industry. Reliability and failure prediction are the major constraints for the estimation of the residual life of the component. In this paper, Artificial intelligence techniques are employed on bipolar junction transistor which provides knowledge of failure mechanism of a component in actual operating conditions such that if it deviates from the actual output, a preventive measure to be taken before serious failure occurs. The end of life has been explored using the design of experiments approach. After calculating lifetime, an expert system has been modeled which predicts the sudden crash of transistor before it actual fails, using various statistical and analytical techniques. The comparison of accuracy has been conducted on all techniques of artificial intelligence and statistical method. The comparison shows that ANFIS is the most accurate technique with an accuracy of 96.65%. A graphical user interface is created which indicates the failure of bipolar junction transistor at various level of inputs.