Smart devices are revolutionizing daily life by making them more intelligent and efficient. In agriculture, sensors are being used to detect soil pH, moisture and nutrients in order to grow high-quality crops. The health of crops is heavily influenced by the quantity and quality of the soil, which can be affected by physical factors such as soil composition and density, as well as chemical factors such as nutrient accessibility. By using the DS18B20 Waterproof Temperature Sensor to measure soil temperature and rain sensors to monitor moisture levels, farmers can gain valuable insights into the health of their soil. Advancement of big data technologies and powerful computers, machine learning is providing new opportunities for research in the field of Agri-technology. By applying machine learning techniques to sensor data, farm management systems are becoming real-time intelligent systems that can provide precise recommendations and insights for farmers to use in their decision-making. In this study, a Recurrent Neural Network (RNN) approach was used to determine the suitability of a soil sample for plant nutrition and soil fertility. Proposed system results and evaluated them with numerous conventional classification algorithms. the RNN higher accuracy of 94% on the real-time dataset. This research has the potential to be highly beneficial for the Agri-tech industry in the near future by providing farmers with valuable recommendations for plant nutrition and soil fertility management.