This paper provides a method for biometric authentication using voice for speaker recognition. In the current day and age, security has become a major concern, only passwords aren’t sufficient to ensure data safety. Hence it has become necessary to incorporate biometric features into security systems. The paper uses Mel-Frequency Cepstral Coefficients (MFCC) as the unique features for speaker as well as speech recognition. Neural Networks (NN) have been used for classification of input voice samples as authenticated or non-authenticated and to identify the word spoken. This method has been implemented using a rover, which uses this biometric authenticated system. The rover gets powered on only when the “Rover Start” command received is from an authenticated user. Once initiated, the rover can execute direction commands such as “Left”, “Right”, “Back”, “Forward”, and “Stop” given by any user.