The majority of studies on breastfeeding consists of descriptive correlational studies identifying the incidence and correlates of breastfeeding. The theory of planned behavior has been shown to yield great predictive power for behavioral goals over which individuals have only limited control such as improving school grades and weight loss. The purpose of this study was to test the "theory of planned behavior" in the prediction of breastfeeding of mothers who delivered vaginally. One hundred mothers who delivered vaginally in one general hospital in Seoul and one general hospital and three private hospitals in Taejeon participated in this study. The instruments used for data collection in this study were developed by the researchers following the guidelines suggested by Ajzen and Fishbein(1980) and Ajzen and Madden(1986). The instruments included measurement of attitude, subjective norm, perceived behavioral control and intention. The collected data were analyzed using descriptive statistics, Pearson product moment correlation, hierachical multiple regression and logistic regression. The results are as follows : 1. Intention to breastfeed correlated significantly with attitude, subjective norm and perceived behavioral control. Both attitude and subjective norm did not make a significant contribution to the prediction of intention, but the addition of perceived behavioral control to the regression equation greatly improved the model's predictive power, increasing the R(2) from .05 to .52. 2. Intention to breastfeed alone had a significant predictive effect on actual breastfeeding, resulting in a regression coefficient of .16(x2=8.60, p<.01), but when perceived behavioral control was added to the equation, intention was not a significant predictive variable and only perceived behavioral control showed significant predictive power on actual breastfeeding, resulting in a regression coefficient of .12(x2=4.68, p<.05). In sum, breastfeeding behavior lent only partial support to the second version of the theory of planned to the second version of the theory of planned behavior, and because perceived behavioral control had a strong effect on intention to breastfeed and actual breastfeeding, it would be desirable to develop nursing intervention programs which focus on strengthening the perceived behavioral control for the promotion of breastfeeding.