Objective: To examine how physical activity and physical mobility are related to obesity in the elderly. Study design: A cross-sectional study of 2558 men and women aged 65 years and older who participated in a population survey in 2012 was conducted in mid-Sweden with an overall response rate of 67%. Methods: Obesity (body mass index >= 30 kg/m(2)) was based on self-reported weight and height, and physical activity and physical mobility on questionnaire data. Chi-squared test and multiple logistic regressions were used as statistical analyses. Results: The overall prevalence of obesity was 19% in women and 15% in men and decreased after the age of 75 years. A strong association between both physical activity and obesity, and physical mobility and obesity was found. The odds for obesity were higher for impaired physical mobility (odds ratio [OR] 2.83, 95% confidence interval [CI] 2.14-3.75) than for physical inactivity (OR 1.63, 95% CI 1.28-2.08) when adjusted for gender, age, socio-economic status and fruit and vegetable intake. However, physical activity was associated with obesity only among elderly with physical mobility but not among those with impaired physical mobility. Conclusion: It is important to focus on making it easier for elderly with physical mobility to become or stay physically active, whereas elderly with impaired physical mobility have a higher prevalence of obesity irrespective of physical activity. (C) 2017 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.