A computer access security system, a reliable way of preventing unauthorized people for accessing, changing or deleting, and stealing the information, needed to be developed and implemented. In the present study, a neural network based system is proposed for computer access security for the issues of preventive security and detection of violation. Two types of data, time intervals between successive keystrokes during password entry through keyboard and voice patterns spoken via a microphone, are considered to deal with a situation of multiple users where each user has a certain password with different length. For each type of data, several multi-layered neural networks are designed and evaluated in terms of recognition accuracy. A committee neural network is formed consisting of six multi-layered neural networks. The committee decision was based on majority voting of the member networks. The committee neural network performance was better than the neural networks trained separately.