ObjectiveTo culturally adapt and examine the psychometric properties of the Chinese (Cantonese) version of the Upper Extremity Functional Index (C-UEFI) in people with chronic stroke.DesignCross-sectional study.SettingsUniversity-affiliated neurorehabilitation research laboratory.ParticipantsThe participants (N = 151) were people with chronic stroke (N = 101) and healthy controls (n = 50).Main outcome measuresWe assessed the C-UEFI, Fugl-Meyer Assessment for Upper Extremity (FMA-UE), Wolf Motor Function Test (WMFT), Six-Minute Walk Test (6MWT), Motor Activity Log (MAL), Activity-Specific Balance Confidence (ABC) scale, Lawton Instrumental Activities of Daily Living (IADL) scale, Survey of Activities and Fear of Falling in the Elderly (SAFFE), Stroke Impact Scale (SIS) and Community Integration Measure (CIM) as outcome measures.ResultsThe C-UEFI items demonstrated good test–retest reliability (intraclass correlation coefficient [ICC]3, 1 = 0.872) and excellent internal consistency (Cronbach's α = 0.922). People with chronic stroke had poorer C-UEFI scores than the healthy controls. The overall C-UEFI mean score of 101 people with stroke was significantly correlated with the mean scores of the FMA-UE, WMFT, MAL, ABC scale, IADL scale, SAFFE, SIS and CIM and the distance covered in the 6MWT. The C-UEFI cut-off score to distinguish between people with chronic stroke and healthy older adults according to upper extremity function was 57.5 out of 59 (sensitivity: 88.1%; specificity: 84%). The C-UEFI had good content validity, with an acceptable fit to the two-factor structure model.ConclusionsThe C-UEFI is reliable and valid for assessing functional recovery of upper extremity activity in Chinese people with chronic stroke.