The objectives of this study were to verify the effectiveness of the Theory of Planned Behavior in predicting exercise intention and exercise behavior, and to examine the determinants of exercise intention and exercise behavior in the sample of middle-aged women. The subjects who participated in this study were 263 middle-aged women. The instruments used for this study was a survey of general characteristics, attitude (18 items), subjective norm (2 items), perceived behavioral control (19 items), intention (3 items), and exercise behavior (7 items and 23 items for each). Analysis of data was done by use of descriptive statistics and correlation analysis with SAS PC program. The hypothetical model based on the Theory of Planned Behavior was tested by use of LISREL 8.12a program. 1) The overall fit of the hypothetical model to the data was good(chi-square=11.76, p=0.03, RMSEA=0.07, standardized RMR= 0.03, GFI=0.99, AGFI=0.94, NFI=0.97, NNFI=0.95). 2) Perceived behavioral control, attitude, and subjective norm were significant determinants of exercise intention, and these variables explained 35% of the total variance of exercise intention. 3) Perceived behavioral control, intention, and attitude were significant determinants of exercise behavior. But, subjective norm was not a significant determinants. These four variables explained 69% of the total variance of exercise behavior. In conclusion, this study shows the applicability of the Theory of Planned Behavior in explaining exercise behavior of middle-aged women, and suggests that health care providers should focus on perceived behavioral control and attitude rather than subjective norm to improve exercise behavior of middle-aged women.