Objective: To establish normative values and a reference equation for the number of steps climbed during the six-minute step test (6MST) in healthy adults, and to assess the reliability of the test and of the equation. Methods: This was a multicenter cross-sectional study involving 468 healthy volunteers (age range: 18-79 years) recruited from the general community in six research laboratories across different regions of Brazil, which is a country with continental dimensions. The 6MST was performed twice (30-min interval), and clinical, demographic, and functional variables were evaluated. An independent sample of 24 volunteers was evaluated to test the reference equation a posteriori. Results: The number of steps had excellent test-retest reliability (intraclass correlation coefficient = 0.96 [95%CI: 0.95-0.97]), and the mean number of steps was 175 ± 45, the number being 14% greater in males than in females. The best performance on the test was correlated with age (r = -0.60), sex (r = 0.28), weight (r = 0.13), height (r = 0.41), BMI (r = -0.22), waist circumference (r = -0.22), thigh circumference (r = 0.15), FVC (r = 0.54), and physical activity level (r = 0.17; p < 0.05 for all). In the regression analysis, age, sex, height, and weight explained 42% of the variability of the 6MST. Normative values were established for the 6MST according to age and sex. There was no difference between the 6MST values from the independent sample and its predicted values (157 ± 29 steps vs. 161 ± 25 steps; p = 0.47; 97% of predicted values). Conclusions: The normative values and the reference equation for the 6MST in this study seem adequate to accurately predict the physical functional performance in adults in Brazil.