Abstract
Anself-governing ATM host has aright to use any bank. There is no security layer is implemented in the ATM card except pin number. It is very costly for the bank to include the fingerprint and Iris scanner. In this paper, we monitor the location of the ATM usage, time taken for the user to accessing the ATM machine, sequence of events processed by the user and expected amount of withdrawal by the user. All these four factors are verified for the authentication purpose of the user along with password. If any of the above said, parameter are differing and then the One Time Password is generated to the User’s Mobile number for further more secure authentication system. In the modification phase, an automation user Internet recognition model is designed to enhance the user comfort and detection of the time span spend by the user in the ATM machine. If due to signal problem of the mobile One Time Password will not be received in that cause secret process is used to private ATM users.