The aim of the study was to develop a non-exercise VO2max predictive model using a physical activity (PA) variable determined by pedometer-determined step counts (SC: steps/day) in active lifestyle men aged between 40 and 65. For this purpose, 69 subjects were used to develop and validate the new model. VO2max was measured using Rockport Fitness Walking Test. Anthropometric variables were measured. PA was measured by a pedometer (Hj-113 Omron, Japan) throughout the day for two consecutive weeks. The results showed a significant correlation between VO2max and independent (anthropometric and daily steps) variables (p>0.05). There was a significant relationship between SC and VO2max (partial correlation coefficient r=0.70, P<0.001). Multiple regression analysis showed that VO2max could be predicted with age, weight, WHR and daily SC (R=0.92, SEE=1.4 ml/kg/min, P<0.001). When the new prediction model was applied to estimate the participants’ VO2max, predicted VO2max and measured VO2max were significantly correlated (r=0.86, P<0.005). It seems that SC is a useful variable for non-exercise prediction of VO2max in middle-aged men. This model may increase the precision of regression models to predict VO2max in health-related studies.