Another effective tool that has proliferated to other spheres is machine learning, including the agricultural one. Nevertheless, there are relatively few research-level studies on the theme of maximizing agricultural output. In this program, an analysis framework based on machine learning algorithms will be developed to improve productivity in agriculture. The proposed approach will take place in three stages to generate pertinent knowledge. We have already conducted data cleaning, data augmentation, data collection, feature selection, feature extraction, and feature construction in the aforementioned stages. In the activities are the suggested optimized technique and the traditional techniques. By utilizing longitudinal and latitudinal data, the suggested method can identify the crop that is most suitable for prediction. The fundamental intention of the model is to ensure that it functions more efficiently as compared to the previous version and achieves the highest level of accuracy. The operations of determining latitude and longitude, forecasting temperature and moisture, assessing the needs for GDD and NPK, and providing structural information are well-defined in the agricultural field and introduce several industrial dimensions. Such treatments will become more precise and useful, respectively, because of the technological development. Other algorithms that follow this method achieve an accuracy of 92%, while the accuracies of ADAM, RSMM, CYEM, and IOTM are 50%, 67%, 78%, and 40% respectively. Further work will aim at improving the quality of the data that has been collected as well as adopting more effective means.