Artificial intelligence (AI), particularly machine learning (ML), has become a key technology in antenna design due to its ability to model nonlinear relationships between antenna dimensions and performance. ML enables accurate prediction of antenna performance parameters based on input dimensions, reducing optimization time and the number of simulations compared to traditional methods. However, some existing studies still face limitations such as small datasets, high prediction errors, or complex and manual data processing procedures. This study proposes a data processing method on Google Colab to rapidly and accurately construct training datasets for machine learning. Based on a dataset of 15,000 samples created by this method, the Gradient Boosting (GB) model is used to predict antenna performance with a mean squared error (MSE) of 0.2914, demonstrating high prediction accuracy and low error, thereby significantly reduce simulation and antenna design optimization time.