The purpose of this study was to investigate whether patients can accurately self-assess their knee passive range of motion (PROM). A picture-based questionnaire for patient self-assessment of knee PROM was developed and posted to patients. The self-assessed PROM from 58 patients was compared with surgeon-assessed PROM using a short-arm goniometer. Agreement between the measurement methods was calculated with the Bland-Altman method. We calculated the sensitivity and specificity of patient-assessed PROM in dichotomously detecting knee motion impairment in both flexion (≤ 100 degrees) and extension (≥ 10-degree flexion contracture). Surgeon- and patient-assessed knee PROM showed a mean difference (95% limits of agreement) of -2.1 degrees (-42.5 to 38.3 degrees) for flexion and -8.1 degrees (-28.8 to 12.7 degrees) for extension. The sensitivity of patient self-assessed PROM in identifying knee flexion and extension impairments was 86 and 100%, respectively, whereas its specificity was 84 and 43%, respectively. Although wide limits of agreement were observed between surgeon- and patient-assessed knee PROM, the picture-based questionnaire for patient assessment of knee ROM was found to be a valid tool for dichotomously detecting knee motion impairment in flexion (≤ 100 degrees). However, the specificity of the questionnaire for detection of knee extension impairments (≥ 10-degree flexion contracture) was low, which limits is practical utility for this purpose.