Monte Carlo simulated SEM images for realistic instrumental conditions are used to evaluate measurement methods for SEM image sharpness. The Monte Carlo simulation of the SEM image is based on a well-developed physical model of electron-solid interaction, which employs Mott’s cross section for elastic electron scattering and dielectric functional approach to electron inelastic scattering with cascade secondary electron production included, a finite element mesh modeling of complex sample topography and a modeling of SEM instrumental conditions (i.e. focus, astigmatism, drift and vibration). A series of simulated SEM images of a realistic sample, gold particles on a carbon substrate, for different instrumental parameters are generated to represent practical images where all instrumental conditions are precisely known and controlled. An estimation of three measurement methods of SEM image sharpness, i.e. FT, CG and DR methods, has then been performed with these simulated images. The responses of image sharpness measurement methods to various instrumental conditions are studied. The calculation shows that all the three methods present similar and reasonable response to focus parameter; their dependences of the measured sharpness on astigmatism coefficient are complicated and CG method presents reasonable sharpness value. For drift and vibration, the situation is more complex because CG/DR methods can be less or more sensitive to vibration coefficient than FT method. Because of the different response behaviors of the three sharpness measurement methods to experimental parameters, we propose to use a mean, simple average or weighted average, of three sharpness values as a proper measure of sharpness.