ImportanceThere is a lack of comprehensive data on the association of the COVID-19 pandemic with the prevalence of physical activity in large-scale data sets.ObjectiveTo investigate long-term trends in physical activity using information from a nationally representative survey covering 2009 to 2021.Design, Setting, and ParticipantsThis general population–based and repeated cross-sectional study was conducted from 2009 to 2021 using the Korea Community Health Survey, a nationally representative survey in South Korea. Data from 2009 to 2021 for 2 748 585 Korean adults were obtained through a nationwide, large-scale, serial study. Data were analyzed from December 2022 through January 2023.ExposureCOVID-19 pandemic onset.Main Outcomes and MeasuresThe trend of sufficient aerobic physical activity was measured by prevalence and mean metabolic equivalent of task (MET) score based on World Health Organization physical activity guidelines and defined as 600 MET-min/wk or greater. The cross-sectional survey included age, sex, body mass index (BMI), region of residence, education level, income level, smoking status, alcohol consumption level, stress status, physical activity level, and history of diabetes, hypertension, and depression.ResultsAmong 2 748 585 Korean adults (738 934 aged 50-64 years [29.1%] and 657 560 aged ≥65 years [25.9%]; 1 178 869 males [46.4%]), the prevalence of sufficient physical activity did not change significantly during the prepandemic period (β difference, 1.0; 95% CI, 0.6 to 1.4). During the pandemic, the prevalence of sufficient physical activity decreased significantly, from 36.0% (95% CI, 35.9% to 36.1%) in 2017 to 2019 to 30.0% (95% CI, 29.8% to 30.2%) in 2020 and 29.7% (95% CI, 29.5% to 29.9%) in 2021. Trends showed decreases in the prevalence of sufficient physical activity among older adults (ages ≥65 years; β difference, −16.4; 95% CI, −17.5 to −15.3) and younger adults (ages 19 to 29 years; β difference, −16.6; 95% CI, −18.1 to −15.0) during the pandemic. In particular, the trend of sufficient physical activity declined during the pandemic in females (β difference, −16.8; 95% CI, −17.6 to −16.0), individuals in urban residences (β difference, −21.2; 95% CI, −22.2 to −20.2), healthy participants (eg, those with normal BMI, 18.5 to 22.9: β difference,−12.5; 95% CI, −13.4 to −11.7), and individuals at increased risk of stress (eg, history of a depressive episode; β difference, −13.7; 95% CI, −19.1 to −8.4). Prevalence trends in mean MET score were similar to those in the main results; total mean MET score decreased from the 2017 to 2019 period (1579.1 MET-min/wk; 95% CI, 1567.5 to 1590.7 MET-min/wk) to the 2020 to 2021 period (1191.9 MET-min/wk; 95% CI, 1182.4 to 1201.4 MET-min/wk.Conclusions and RelevanceThis cross-sectional study found that the national prevalence of physical activity was stable or consistent before the pandemic period, with a marked decrease during the pandemic, particularly among healthy individuals and subgroups at increased risk of negative outcomes, including older adults, females, urban residents, and those with depressive episodes. Future studies may be needed to evaluate the association between the COVID-19 pandemic and changes in physical activity.