Astrology is a very ancient and traditional method of prediction that increases the interest of people continuously. The globe today, there are no common guidelines or principles for astrological prediction. Rather than setting universal principles and criteria for astrological prediction, astrologers focus on providing high-quality services to individuals but there is no guarantee of accuracy. Machine learning is providing the best result for analysis and prediction on many applications by the learning of computers. Prediction and classification make it possible for any learner to work on large, noisy, and complex datasets. The main motive of the paper is to introduce a scientific approach that reduces the drawback of the traditional approach and indicates the universal rules of prediction and proves the validity of astrology by the three classification techniques, Naïve Bayes, Logistic-R, and J48. It is a part of supervision learning that operates with cross-validation 10,12, and 14fold for calculating the terms 1) correctly classified instances (CCI), erroneously categorized instances (ECI), Mean absolute error (MAE), Root mean squared error (RMSE), and Relative absolute error (RAE). 2) True Positive Rate, False Positive Rate, Precision, and F-Measure values. 3) The MCC, ROC, and PRC area values. 4) To calculate the average weight of the three-class label professor, businessman, and doctor in terms of true positive rate, false-positive rate, precision, F-measure, PRC, and ROC area, 5) finally, we calculated the accuracy of each classification technique and compare which provide the better result. For this, we have collected the date of birth, place of birth, and time of birth of 100 persons who belong to different professions. 40 data of professors, 30 data of businessmen, and 30 data of doctors, prepare the horoscope of an individual with the help of software. For analysis, we create the datasheet in .csv format and apply this data sheet in the weka tool to check various parameters and the accuracy percentage of each classifier.