This paper is concerned with application of interconnected learning automata to learning coordinated control (LCC) of power systems. The interconnected learning automata with variable subsets of control actions are used to optimize parameters of distributed local controllers to improve power system stability based on an integrated performance index. Simulation studies have been carried out based on a detailed mathematical model of a four-machine power system. Results obtained from simulation studies are presented to show a potential of applications of the method to control of large-scale systems.