Charge eXchange Recombination Spectroscopy (CXRS) systems have been successfully applied on EAST since 2014. However, geometrical effects will deteriorate the diagnostic locality and introduce uncertainties into the measurement. A simplified spatial average model for the average of effective integral paths and an upgraded simulation of spectra (SOS) code are both used to evaluate these geometrical effects. The results show that spatial averaging effect with one beam can introduce up to ˜10% relative error in the ion temperature and velocity measurement at ITB region. Simultaneous injection of two beams can produce ˜20% uncertainties in core CXRS with single gaussian description of the spectra. To reduce the error, the spatial averaged position ρav calculated from the simplified model and a double gaussians fitting with a constrained intensity ratio model are applied in the CXRS data analysis, and the fitting results show a good consistency with the input profiles.