Physical activity (PA) plays an important role in maintaining a healthy weight. To develop targeted strategies that encourage physical activity, knowledge of associations between intensity of physical activity (PA) levels and body mass index (BMI) is essential. We aimed to examine the relationship between sufficient and vigorous PA and BMI distribution among adults in northwest China using quantile regression. We conducted a cross-sectional study in Shaanxi Province in 2013, using proportional probability sampling. BMI was calculated using measured height and weight. The Global Physical Activity Questionnaire was used to define and measure sufficient and vigorous PA. Associations of sufficient/vigorous PA and BMI were modelled using quantile regression. Mean BMI was 24.18 ± 3.51 and BMI distribution with age showed an inverse U shape. A total 9045 (88.97%) participants demonstrated sufficient PA and 3119 (30.68%) reported vigorous PA. After adjusting for relevant sociodemographic, dietary, and lifestyle parameters in quantile regression modelling, sufficient PA was positively associated with BMI score distribution from the 1st to 30th quantile, with β from 0.32 (95% confidence interval (CI): 0.07 to 0.63) to 0.85 (95% CI: 0.40 to 1.19). Vigorous PA was negatively associated with BMI score distribution from the 30th to 93th quantiles, with β from -0.18 (95% CI: -0.31 to -0.02) to -0.81 (95% CI: -1.10 to -0.45). Sufficient PA was positively associated with underweight and normal weight whereas vigorous PA was negatively associated with overweight and obesity.