Objective: To evaluate the relationship between self-reported and performance-based measures of functioning in rheumatoid arthritis (RA), knee osteoarthritis (OA), and fibromyalgia (FM), and the influence of pain and fatigue. Method: Self-reported functioning was assessed by the Stanford Health Assessment Questionnaire, Fibromyalgia Impact Questionnaire, and Knee injury and Osteoarthritis Outcome Score. Performance-based measures of task-related physical activity included grip strength and Six-Minute Walk Test (6MWT). Assessment of Motor and Process Skills (AMPS) was used to obtain performance-based measures of activities of daily living (ADL) ability. Pain and fatigue were assessed by 100 mm visual analogue scales. Spearman’s rho correlation and regression modelling were applied. Results: Correlations between self-reported functioning and performance-based measures of ADL ability were weak to moderate, and strongest in OA (r = 0.57, p = 0.002), and AMPS ADL ability measures did not enter regression models as explanatory factors for self-reported functioning. Correlations between AMPS ADL ability measures and measures of task-related physical activity were weak, except for a strong correlation between AMPS ADL motor ability and 6MWT in OA (r = 0.63, p = 0.000). The 6MWT was the only performance-based test explaining variance in AMPS motor ability (OA = 42%; FM = 11%). Pain explained variance in self-reported ability and contributed to variance in AMPS ADL motor ability measures in OA. Conclusion: Self-reported and observed measures of functioning assess partly different aspects of functioning, and both approaches may therefore be relevant in a structured assessment of patients with musculoskeletal disorders.