OBJECTIVE: To assess the relationship between walking for exercise and symptomatic and structural disease progression in individuals with knee osteoarthritis (OA). METHODS: We assessed a nested cohort of participants age 50 years or older within the Osteoarthritis Initiative, a community-based observational study in which subjects were enrolled between 2004 and 2006. We focused on 4 dichotomous outcomes from baseline to the 48-month visit, involving determination of the frequency of knee pain and radiographic severity of knee OA on posteroanterior semiflexed knee radiographs. The outcomes assessed included 1) new frequent knee pain, 2) worsening of radiographic severity of knee OA based on the Kellgren/Lawrence grade, 3) progression of medial joint space narrowing, and 4) improved frequent knee pain. We used a modified version of the Historical Physical Activity Survey Instrument to ascertain those subjects who reported walking for exercise after age 50 years. The survey was administered at the 96-month visit (2012-2014). RESULTS: Of 1,212 participants with knee OA, 45% were male and 73% reported walking for exercise. The mean ± SD age was 63.2 ± 7.9 years, and the mean ± SD body mass index was 29.4 ± 4.6 kg/m . The likelihood of new frequent knee pain was reduced in participants with knee OA who walked for exercise as compared to those who were non-walkers (odds ratio [OR] 0.6, 95% confidence interval [95% CI] 0.4-0.8), and progression of medial joint space narrowing was less common in walkers compared to non-walkers (OR 0.8, 95% CI 0.6-1.0). CONCLUSION: In individuals with knee OA who were age 50 years or older, walking for exercise was associated with less frequent development of knee pain. These findings support the notion that walking for exercise should be encouraged for people with knee OA. Furthermore, we offer a proof of concept that walking for exercise could be disease modifying, which warrants further study.