Harvesting plants for their medicinal uses creates an intersection between agriculture and medicine. The final product’s value and efficacy are of utmost importance. These attributes depend on certain conditions which are difficult to control due to the limitations of conventional farming techniques. The process of soil evaluation for agriculture also poses challenges. These evaluations can be time consuming and costly while also producing results that are often contradictory. This article attempts a different angle. It focuses on the convergence of Machine Learning (ML), Long Range (LoRa) communication and Wireless Underground Sensor Networks (WUSNs) for the first time in monitoring soil conditions of crops (medicinal plants) in real time and on a fine-grained scale. The proposed system uses buried WUSN nodes to collect granular data on critical soil parameters to be monitored continuously: moisture, pH, temperature, and essential macronutrients Nitrogen, Phosphorus, and Potassium (NPK). This data is relayed over long distances using low-power communication, a real time feature critical for many applications, to the cloud for data processing. An ML model then assesses the soil and offers predictions on the health of the crops (plants) and stresses that will be or are affecting the crop. The model will also offer recommendations to alleviate the impact of stress and enhance the effectiveness of the phytochemicals. This model offers a robust and scalable solution.