Aim To cross‐culturally adapt and validate the Comprehensive Frailty Assessment Instrument (CFAI) among Chinese community‐dwelling older adults. Methods The Chinese CFAI was developed through forward–backward translations. An urban sample of 1235 community‐dwelling older adults received face‐to‐face interviews to examine the validity (construct validity and criterion validity) and reliability (internal consistency and test–retest reliability). Results The Chinese CFAI achieved semantic and idiomatic equivalence, and showed acceptable reliability and an expected factor structure, except for the social support domain. The exploratory factor analysis extracted five factors explaining 53.8% of the total variance of frailty. The confirmatory factor analysis showed that the data fit well to the second‐order factor theoretical model, with a root mean square error of approximation of 0.05, Tucker‐Lewis Index of 0.93 and Comparative Fit Index of 0.95. The receiver operating characteristic analysis presented an acceptable criterion validity using the Rockwood Frailty Index as an external criterion (area under the curve 0.80), with balanced sensitivity (65.31%) and specificity (81.19%) at the optimal 39‐point frailty cut‐off of the CFAI. Conclusions The Chinese CFAI has good validity and reliability as a practical frailty measure in Chinese community‐dwelling older adults.