The present study investigated the use of the accelerometer-determined physical activity (PA) variables as the objective PA variables for estimating VO(2max) in Japanese adult men. One hundred and twenty-seven Japanese adult men aged from 20 to 69 years were recruited as subjects of the present study. Maximal oxygen uptake VO(2max) was measured with a maximal incremental test on a bicycle ergometer. Daily step counts (SC) and the amount spent in moderate to vigorous PA (MVPA) and vigorous PA (VPA) were measured using accelerometer-based activity monitors worn at the waist for seven consecutive days. The non-exercise models were derived using hierarchical linear regression analysis, and cross-validated using two separate cross-validation procedures. SC, MVPA, and VPA were significantly related to VO(2max) (partial correlation coefficient r = 0.58, r = 0.42, and r = 0.51, respectively) after adjusting for age. Two models were developed by multiple regression to estimate VO(2max) using data of age, SC, VPA, and either BMI (the coefficient of determination (R(2)) = 0.71, standard error of estimate (SEE) = 4.2 ml kg(-1) min(-1)), or waist circumference (R(2) = 0.74, SEE = 3.9 ml kg(-1) min(-1)). All regression models demonstrated a high level of cross-validity supported by the minor shrinkage of R(2) and increment of SEE in the PRESS procedure, and by small constant errors for subgroups of age, SC, and VO(2max). This study demonstrated that combining SC with VPA, but not with MVPA, was useful in predicting VO(2max) variance and improved the ability of the regression models to accurately predict VO(2max).