A new Model-order Reduction (MoR) procedure, that combines Galerkin projection method with H₂-norm minimization technique, is developed for linear continuous-time systems. The proposed algorithm, stated as the solution of a set of Linear-Matrix-Inequality (LMI) conditions, is subsequently utilized in MoR of nonlinear systems with stability preservation. The developed algorithms are dedicated to stable MIMO linear systems and stable MIMO nonlinear systems having repetitive nonlinearities. The obtained results are compared to the classical balanced truncation algorithm, the Hankel MoR method, and the recent incremental balanced truncation procedure. The designs are illustrated through many examples including a case of a nonlinear high-order electrical circuit.