Background: The physical activity (PA) level of individuals with knee osteoarthritis is lower than in the general population. International recommendations recommend a non-pharmacological intervention including a self-management education program, weight loss and an adapted exercise program. However, we have no scale assessing the perceived barriers to and facilitators of PA in this population. Objective: We constructed and validated a self-administered questionnaire assessing perceived barriers to and facilitators of regular practice of PA in people with knee osteoarthritis. Methods: Semi-structured interviews identified 24 barriers and facilitators. We developed a 24-item questionnaire, Evaluation of the Perception of Physical Activity (EPPA) that was completed by 548 individuals with knee osteoarthritis, to assess acceptability, construct validity, internal consistency and convergent validity. Participants also completed the Knee Osteoarthritis Fears and Beliefs Questionnaire (KOFBeQ), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and International Physical Activity Questionnaire (IPAQ). Reliability and sensitivity to change were evaluated in a second group of 168 people with knee osteoarthritis at a 3-week spa therapy resort. Results: Factorial analysis identified 17 items grouped into 4 subscales (Barriers, Facilitators, Motivation and Beliefs). The internal consistency was good for Barriers, Facilitators and Motivation subscales (Cronbach alpha > 0.70) and intermediate for the Beliefs subscale (Cronbach alpha = 0.64). The EPPA subscale scores were significantly correlated with KOFBeQ and WOMAC scores but not associated with IPAQ physical activity level. Reliability was good for all subscales, with intraclass correlation coefficients > 0.60. A sensitivity to change was found for only the Beliefs subscale, with a moderate effect size. Conclusions: The EPPA questionnaire has good psychometric properties and can help guide the management of knee osteoarthritis. It can be used in research for evaluating the perception of physical activity.