Humanity cannot continue to exist without agriculture. Many people around the world rely heavily on agriculture for their primary source of income. It also offers a wide variety of job opportunities to people. Many farmers long for the days of traditional farming, despite the fact that it offers very little financial reward in present marketplace. The future enlargement and wealth of the American economy hinge on the health of the agricultural sector and related businesses. Selecting optimal crops and establishing necessary infrastructure can improve crop yields for agribusinesses. Predictions in agriculture take into account a wide range of variables, including climate, soil-fertility, water-availability, water-quality, crop-pricing, and more. Predicting agricultural output from variables like location, climate, and harvest season is impossible without the use of machine learning (ML). Predicting market pricing, organizing import and export activities, and minimizing the social cost of agricultural losses are all facilitated by accurate estimates of crop yields throughout the budding time, which is beneficial for both policymakers and farmers. Farmers can use this tool to better determine what crops would do best on their property. In this research, we provide a ML strategy for forecasting crop yields. The data set for an experiment may include crop data. The relief method is then used to choose features. The linear discriminant investigation technique is used to extract the features. Particle swarm optimization-support vector machine (PSO-SVM), K-nearest neighbor (KNN), and random forest (RF) are some of the ML predictors utilized for categorization.