Objective To assess the psychometric potential of the Montreal Cognitive Assessment Scale-Beijing (MoCA-BJ) as a screening instrument for mild cognitive impairment (MCI) in older adults in Wuhan communities of central China. Methods MoCA-BJ and Mini-Mental State Examination (MMSE) were adopted to assess the MCI of 381 older adults from 13 communities in Wuhan in 2015. Confirmatory factor analysis was conducted to evaluate the construct validity of MoCA-BJ, and the relationship between all aspects of cognitive function and MoCA different dimensions. Results MoCA-BJ had acceptable reliability (w=0.76), and MoCA-BJ and MMSE estimation results were highly correlated (r=0.73, P＜0.01). By comparing three measurement models through confirmatory factor analysis, we found that the MoCA-BJ scale had two factors (F1: visual space executive function, F2: memory-based other cognitive functions) in model 3, fit degree of which was higher than model 1 by one factor, and there was a statistically significant difference in the number of factors between model 1 and model 3 (χ2dif=8.73,P＜0.01). Conclusions The MoCA-BJ has two underlying factors that respectively represent two highly correlated but distinct factors, cognition and visual-spatial. Uninformative items should be revised with culturally sensitive items and the cut-off point for mild impairment should also be altered.