California Bearing Ratio (CBR) value has been widely used to evaluate pavement foundation characteristics. To minimize the effect of human errors, cost, and time for selecting soil subgrade soil for roads, the CBR value can be developed using regression techniques by performing numerous CBR and physical tests considering different soil types. The main objective of this study was to examine the correlation of CBR value with soil properties of road subgrade. Twenty-seven specimens were obtained from twenty different provinces in Cambodia. The basic properties tests (Sieve analysis, Atterberg limit, compaction test, and CBR test) were conducted. From the test result, multiple linear regression was adopted to correlate the prediction model of the CBR. Based on the current study, it was found that the prediction model with the function of gravel, sand, fine, plastic index, maximum dry density, and optimum water content provided a better coefficient of determination (R2) for both study and validating data, which was about 0.9215 and 0.8348, respectively. However, another model is preferable practically since it relates only to the sieve analysis parameter. That model also has better R2 for training (R2 = 0.8901) and validating data (R2 = 0.6969). Therefore, that model should commonly be used for the primary check of the soil in the field due to human effect, costly and time-consuming.