Despite agriculture’s importance to India’s economy, the proposed approach still haven’t figured out how to profitably utilize huge agricultural land. The fundamental reason is a lack of knowledge about soil. Different types of soil have their own peculiarities. Soil property testing is so crucial. There may be a lot of tools for soil analysis out there, but not all of them are trustworthy, and farmers still have to deal with the trouble of getting dirt from labs. But getting all soil types tested in a timely manner at the lab is no easy feat. The current soil analysis tools also lack translations for regional languages. Consequently, the farmer requires a tool capable of performing soil analysis. Preprocessing the data, choosing features, and training the model should be the top priorities. Before using any machine learning algorithm on a dataset, preprocessing is necessary. It is common practice to collect raw data from multiple sources. The proposed approach use statistical correlation coefficient and information gain in feature selection. It is required to select features before training DE-ANN models. The suggested method performs noticeably better than the two front-runners, DE and ANN. Using the method resulted in a 97.65% increase in accuracy.