In this study, the ratio of compaction was determined by sand cone and nuclear test methods performed in the same place. The normal distribution fitness of the compaction percentiles was tested by the Anderson-Darling test and normal distribution was normalized by subjecting the unsuitable Water Content values to Jonson transformation. Findings the results obtained with the nuclear method reveal that both methods are consistent with each other in terms of mean data, with lower and more variables in the Sand Cone method. It is recommended to test the consistency of the results obtained by the nuclear method. Also, because of the rapid results, statistical model has been investigated for adapting the results of nuclear method to the sand cone method. For this purpose, regression analysis and artificial neural network models were investigated; artificial neural networks were seen to provide successful predictions inside of the found models.