In practice, many industrial control loops inevitably include nonlinearites, so the estimates of the control performance may not be correct. Firstly, the existence of the minimum variance performance lower bound(MVPLB) for a class of nonlinear systems is analyzed and the relation between the MVPLM and disturbance terms is determined. Then, the model of closed-loop system is identified by using orthogonal least square algorithm. Based on the achieved model, the contribution to the output variance due to the uncertainties in most recent ahead disturbance terms is calculated according to ANOVA-like decomposition formula, so that the control performance of the nonlinear system is obtained. Finally, simulation results show the effectiveness and feasibility of the proposed algorithm.