Recently, for estimation of cardiorespiratory fitness, the authors have developed a useful 12-min submaximal treadmill walk test (STWT) in which the intensity is regulated by ratings of perceived exertion (RPE). The purpose of this study was to examine the validity of the STWT for estimation of training-induced changes in cardiorespiratory fitness. The subjects were 9 obese women aged 42.0±4.5 yr, who participated in our exercise program, which was based on 75 min of step exercise 3 times a week for a period of 4 months. All subjects were asked to walk for 12 min at an intensity corresponding to 13 on the Borg scale of RPE, and their VO_<2peak> and VO_<2AT> were measured on the treadmi11 by means of a graded exercise test (GXT). The STWT and the GXT were readministered 4 months after completion of the exercise program. The correlations of 12-min walk distance with VO_<2peak> or V0_<2AT> were found to be statistica11y significant (r=0.76 and 0.70, respectively) when pre-and post-training data were pooled and analyzed. The standard error of estimation (SEE) of the linear regression equation for estimating VO_<2peak> and VO_<2AT> from the 12-min walk distance were 3.01ml/kg/min and l.85ml/kg/min, respectively, both of which were relatively low. After the 4-month exercise program, the coefficient of correlation between the changes in 12-min walk distance and those in VO_<2AT> was statistica11y significant (r=0.78). From heart rate (HR) data measured during the STWT and the GXT, the %HR_ of the STWT was not significantly different from the %HR_ corresponding to RPE13 during the GXT before and after exercise training. From this limited sample size, it was concluded that the STWT at a self-selected intensity intended to correspond to RPE13 is a valid method for estimating changes in the cardiorespiratory fitness of obese women.