Cardiovascular disease (CVD) is a leading cause of death and disability in the Republic of Korea (hereafter “Korea”) and around the world. CVD is known to be associated with a variety of risk factors, and modification of modifiable risk factors is one of CVD prevention strategies. Risk prediction is a critical feature of disease prevention. Existing CVD prevention and management guidelines emphasize the importance of accurate risk prediction and subsequent decision making. However, there is no national clinical guideline for CVD risk estimation in Korea. Thus, this dissertation research aimed to explore the current use and application of CVD risk prediction in Korea and to examine the predictive accuracy of an existing CVD risk prediction model and to investigate the best fitting CVD risk prediction model in a Korean population based-cohort data. Existing literature was reviewed to explore the current use of the Framingham Risk Score (FRS) in Korea, and the National Health Insurance Service-Health Screening Cohort (NHIS-HELAS) was used to examine the predictive accuracy of FRS in Korea. A datamining approach, random survival forest, was adopted to find the best fitting model for Korean adults. The application of the FRS in Korea was categorized into three groups according to the purpose of use: to identify subjects’ CVD risk score, to present a standardized method for CVD risk prediction and to evaluate model performance. To the best of our knowledge, the FRS had not been validated with the general population of Korea. In this study, the FRS has adequate predictive ability to estimate 10-year CVD risk in NHIS-HEALS. The performance of the random survival forest-based models were comparable with the FRS although the random forest-based models demonstrated better agreement between predicted and observed CVD events than the FRS for both men and women in NHIS-HEALS. In conclusion, the FRS is in an early stage of validation in Korea and the FRS performed well in a Korean population-based cohort. The application of a data mining approach slightly improved the predictive accuracy of CVD risk prediction.