Soil quality and fertility are critical determinants of agricultural productivity, governed by the availability of essential nutrients such as nitrogen, potassium, and other key elements. This study presents an advanced machine learning approach for classifying soil quality using a refined multilayer perceptron (MLP) model. The model demonstrates high classification accuracy and strong generalization performance across diverse soil data. Through architectural and functional enhancements, the proposed method offers a robust and scalable solution for evaluating soil conditions. By facilitating timely and data-driven decision-making in precision agriculture, this work contributes to improved soil management practices and the potential for increased crop yields.