Nowadays, agriculture faces several challenges in ensuring food safety. Water scarcity is one of the main challenges facing farmers in the rainfed agriculture sector, especially during the summer, leading to severe economic and farm losses. Internet of Things (IoT) has recently become a potentially revolutionary approach in smart farming that provides many innovative applications. In this research, we suggest an Edge-IoTCloud platform based on a deep learning methodology for monitoring and predicting farmers’ ability to satisfy crop water demands when there is insufficient rainfall. The smart farming system allows collecting data about such important physical phenomena as soil moisture, air temperature, air humidity, water level, water flow, and luminous intensity. The latter is required for reliable and cost-efficient irrigation solutions that will be utilized to compute the necessary water quantity using Rawls and Turq formulas. Cloud services have been chosen for storing and processing significant amounts of data generated by sensors to produce a learning model that will be a basis for predicting future measurements using artificial intelligence and DL techniques. The preliminary results revelated that our proposal is a good starting point for developing low-cost smart farming for smallholder farmers to help them make better decisions.