Limited water resources and the increasing effects of climate change make the sustainable use of water in agriculture a must. Precision irrigation techniques maximize water efficiency in this economic sector; however, these techniques require accurate management of crop water stress to be applied. Traditionally, the identification of crop water stress is a labor-intensive manual process performed with specific machinery (pressure chamber) by experts. Recently, an Internet of Things (IoT) system based on a Wireless Sensor Network (WSN) of leaf-turgor pressure sensors and machine learning models has enabled the digital transformation of this process for citrus crops. In this work, this approach was replicated for plum crops; however, it initially proved ineffective. Consequently, further research was undertaken, involving diverse plum varieties, additional dataset features, and novel pre-processing techniques to enhance the performance of the learning models. This effort resulted in achieving a 90% F1-score in identifying water stress in plum crops.