A new automatic direct digital control strategy for nuclear power reactors is presented. It is based on a simple control logic of comparison between the available time (the time for the error signal to disappear) and the required time (the time for the time derivative to match that of the target trend). The method aims to control the system to an acceptable state within a minimum time under a number of restraints. The control capability of the method is shown for two typical transients. This method is generally applicable to process control in which time-optimal control based on the maximum principle is sought. Though the basic idea follows a heuristic approach, the control strategy is expected to lend itself to robustness against structured perturbations in the system because it is independent of variation in system dynamics.