Abstract Objectives There are problems with the fatigue measures currently used in depressed patients. The Multidimensional Fatigue Inventory (MFI-20) covering general fatigue, physical fatigue, mental fatigue, reduced activity and reduced motivation has been widely used in patients with cancer and chronic fatigue syndrome. To address the multidimensional nature of fatigue, we examined the validity and reliability of a Chinese version of the MFI-20 in major depressive disorder (MDD). Methods Data were derived from a randomized controlled trial of acupuncture in 137 patients with partially remitted MDD. The test–retest reliability, internal consistency, construct and concurrent validity and sensitivity to change of the MFI-20 were analyzed. Results The MFI-20 was found to have good internal consistency (Cronbach's alpha=0.89) and 1-week test–retest reliability (Pearson correlation of the total score=0.73). Factor analysis showed 5 factors, but the factor structure was different from that in medical conditions. The 2 most prominent factors, explaining 46% of the total variance, were both associated with physical and mental energy but different in directions. There were adequate concurrent validity and sensitivity to change as evidenced by the significant correlations between the MFI-20 scores and depressive and anxiety symptoms, general health and quality of life. Conclusion The Chinese MFI-20 is a valid and reliable instrument for the assessment of fatigue in MDD patients with residual symptoms. The construct of fatigue in MDD seems to be different from that in medical conditions. Further studies are needed to examine the MFI-20 in MDD patients from other cultures.