This paper presents a heuristic method for minimization of incompletely specified finite state machine with unspecified values of output variables. The proposed method is based on two states merging. In this method, such optimization criteria as the power consumption and possibility of merging other states are taken into account already at the stage of minimizing internal states. In addition to reduction of the finite state machine (FSM) states, the method also allows reducing the number of FSM transitions and FSM input variables. Experimental results for various styles of state assignment are presented. The results show that this approach to minimization of FSM in most of cases is more effective than classical methods in respect of power consumption.