Soil is one of the most important resources that we must carefully manage. The fertility of the soil affects crop output by around 60%. As a result, regular soil testing is required. Many errors are committed when fertilizing and planting. To ascertain the pH level and nutritional content of the soil, soil tests are utilized. Therefore, deep learning (DL) is important since it has tools for supporting decisions related to cotton crops, as well as tools for supporting decisions related to crop yield prediction. Here, we must use sensors to test the N, P, and K values. The LSTM Model training was carried out using a deep learning algorithm., with a focus on crop yield sequence prediction issues. The findings collected demonstrate that the classification of cotton crop yield using the LSTM algorithm is equally accurate. Then, for cotton that is irrigated, the fertilizer dose is 100:50:50(NPK)kg/ha; for cotton hybrids that are fed by rain, it is 80:40:40(NPK)kg/ha.