Accurate crop classification is crucial for agricultural monitoring and decision-making. Remote sensing's ultimate goal is the precise extraction and classification of crops. Based on a cloud platform, the study area of Guntur district, Andhra Pradesh India, presents a multi-crop classification approach using Sentinel-2 satellite imagery and machine learning techniques. The study area encompasses a diverse agricultural region with three major crop types. After pre-processing, spectral and textural features were extracted. It compares the traditional four machine learning algorithms employed, adding the NDVI, NDBI, MNDWI, and BSI vegetation indices for multi-crop classification enhances accuracy, and offers diverse and complementary information. The overall classification accuracy achieved 95%, with individual crop accuracies ranging from 85 to 96%. The scalable and simple classification method proposed in this research gives full play to the advantages of cloud platforms in data and operation, and the traditional machine learning compared with other algorithms can effectively improve the classification accuracy, and individual areas of crop production are calculated. The results underscored the reliability of GEE-based crop mapping in the region, demonstrating a satisfactory level of classification accuracy, surpassing 97% across distinct time intervals in overall accuracy values, Kappa coefficients, and F1-Score.