Few researchers develop exercise and fitness standards that are individualized for cardiac patients. The majority of researchers have collected exercise data on cardiac patients and calculated mean values for their participants. Exercise prescriptions, tests and programs are then provided to the participants based upon population or sample means. Often mean values are unrealistic and unattainable for cardiac patients. The purpose of this study was to utilize regression analysis to develop a model for estimating individualized treadmill times for cardiac patients and then cross-validate the model to determine its' accuracy. All participants (N=81) were diagnosed with heart disease prior to treadmill testing. Sixty participants were randomly assigned to the validation sample. Physical characteristics (mean±sd) of the 42 male and 18 female participants in the validation sample were: age 53.4±10.7 yr, 54.0±10.8 yr; Body Mass Index (BMI) 27.4±4.1, 28.4±5.7; treadmill time 10.8±2.5 min, 8.0±2.4 min, respectively. Physical characteristics (mean±sd) of the 17 male and 4 female participants in the cross-validation sample were: age 55.9±8.3 yr, 57.3±7.4 yr; BMI 27.8±3.9, 27.7±1.4; treadmill time 9.7±2.4 min, 7.6±1.3 min, respectively. Relative VO2max was estimated from BMI, age, gender, and current exercise history (Jackson et al., 1990). Each participant performed the Bruce treadmill protocol, a well-established protocol used for determining exercise capacity and fitness level in cardiac patients. The first step in the data analysis was to develop a regression model estimating treadmill duration from estimated VO2max. Regression analysis revealed a linear model best fit the data. Quadratic and cubic regression models increased the treadmill time R2 by 1% or less. Regression analysis revealed that estimated VO2max was a significant (p<.05) predictor of treadmill time for cardiac patients. The following regression model (0.215 * estimated VO2max) + 4.571 = time(min) was derived from the validation sample. The accuracy of the derived model was then determined on the cross-validation sample from the following standard error of estimate (SEE) equation: SEE = √S(y-y')2/(n-1). The derived model was found to be accurate (R = .470, SEE = 2.15 min) for estimating treadmill duration times. In conclusion, results of this study revealed the derived model is accurate for estimating treadmill duration times for cardiac patients performing the Bruce treadmill protocol. Predicted treadmill times from the derived model can be used for setting individualized treadmill duration standards for cardiac patients. Keyword(s): adult fitness, measurement/evaluation, medical/medical care