This article explores the use of hybrid FRP Concrete-Steel columns with an elliptical cross section in construction, utilizing machine learning techniques to accurately estimate their performance. The study concludes that the CATB and XGB models offer the highest accuracy and robustness for this type of dataset.