This paper proposes novel technique for multi – machine power system stabilizer using an adaptive Neuro- Fuzzy inference system based on Error reference model (ERANFIS). The adaptive Neuro- Fuzzy inference system based on Error reference power system stabilizer (RNFIS PSS) It utilizes a first-order Sugeno fuzzy logic controller (FLC), the enrollment and outcome function of that is additionally configured online using a neural network accordance with the gradient descent training algorithm based on the reference error model. The simulated power system consists of single – machine – infinite – bus and multi – machine power system. The proposed ERANFIS PSS are designed for each machine. The simulation results use Matlab and S-function technique for various test, and it shows that the proposed stabilizer (ERANFIS PSS) has better performance than that of conventional power system stability (CPSS).