The model predictive controller exhibits excellent constraint and non-linearity handling capabilities. But still, it is less preferred as a centralized controller in the nonlinear standalone micro-grids for two reasons. One is the computational effort involved in predicting the micro-grid response with a nonlinear micro-grid model. The other one is the involvement of more decision variables in the MPC problem. This paper uses the approximated linear model for the forced response prediction to address the first drawback and solves the nonlinear micro-grid model for natural response prediction. Continuous-time Kautz functions are used in the formulation of the MPC problem to address the second drawback. The Kautz functions approximate the control trajectory of each of the inputs within the control horizon. The approximation decreases the decision variables in the MPC control problem. While doing so, the performance capabilities of the controller can still be conserved. © 2020 IEEE