The Classifying stars based on their five inherent properties—spectral class, temperature, luminosity, radius, and absolute magnitude—is a crucial part of astrophysical research. The following examines applying machine learning techniques to a comprehensive analysis of star classification, utilizing the scikit-learn library’s powerful capabilities. The technique consists of exploratory information analysis (EDA), data collection and arrangement, design, modeling, and meticulous optimization. Among the models that will be compared are Calculated Relapse, Decision Trees, Back Vector Machine and Random Forest. Accuracy, which is one of the most important performance metrics measured by perception, is a key performance measurement. However, accuracy is not the only important metric-review score and F1 score are also equally important metrics. Finally, a significance analysis is presented to try to explain which features are most important in making decisions. The results of the analysis show that there is both a good amount of classification accuracy and a great degree of understanding of the real world applications of models for star classification will be and their future direction in astrophysics research. In this study, the first step towards the accurate star classification model is data collection from an open GitHub store. Then, careful cleaning and standardization takes place, and we apply Exploratory Information Examination (EDA) to find the most important features and transform identification with one-hot encoding, which is not presented in this study. After uniting hyperparameters optimization, the model achieved 95% accuracy with temperature and luminosity as the significant features. In conclusion, the review found that Random Forest outperformed the other three models in terms of accuracy and feature significance, improving star classification in astronomy.