INTRODUCTION: The Functional Independence Measure (FIM) is a validated, objective assessment of functional status. It is widely used in rehabilitation centres but may not be practical for all patients due to time and/or personnel constraints. Studies show positive and negative agreement on self-reported FIM scores for patients with spinal cord injuries and amputees. We tested the validity of the self-reported FIM motor score among stroke patients. METHODS: We conducted a prospective double-blind comparative study of patient self-reporting against multidisciplinary assessment, using the standard FIM algorithm. All eligible stroke patients (n = 47) admitted to our rehabilitation centre were included. 33 patients were included in the final analysis. RESULTS: There was substantial agreement on overall FIM motor score between patient self-reporting and multidisciplinary assessment (intraclass correlation coefficient [ICC] 0.651, 95% confidence interval 0.404-0.811). The scores of individual motor items also showed fair-to-good agreement (ICC range 0.431-0.618), except for eating, grooming, bathing and dressing of the lower body (ICC < 0.400). CONCLUSION There was no FIM assessment for 14 (29.8%) patients, highlighting the need for alternative assessment tools. Discrepancies in some scores could be due to patients' ignorance of their own limitations or feelings of embarrassment about reporting. Our results may not be valid for patient populations with cognitive or communication deficits. However, a modest agreement between patient self-reporting and multidisciplinary assessment of FIM motor score was demonstrated. Although patients tend to overrate their performance, self-reported FIM motor scores could be an alternative in situations where multidisciplinary FIM assessment is difficult.