This paper presents a low-cost Internet of Things (IoT) system for automated alternate wetting and drying (AWD) irrigation in paddy fields of the Red River Delta, Vietnam. The system integrates domestically fabricated rainfall, soil-moisture, weather, and ultrasonic water-level sensors with ESP32 gateways, 4G LTE communication, and the ThingsBoard platform. A 15-min control loop selects crop-stage water-level thresholds, suppresses irrigation when 24-h forecast rainfall exceeds 20 mm, and enforces valve and pump safety limits. The system was evaluated on four IoT-managed plots and one conventional continuously flooded plot over the spring and summer rice seasons of 2025. Mean irrigation-water use was 1,477.6 m³/ha in spring and 798.8 m³/ha in summer, representing savings of 33.1% and 33.4%, respectively, relative to the field control. Grain yield reached 8.97 t/ha in spring and 6.72 t/ha in summer, approximately 20% above the corresponding control yields. Water levels remained in the preferred AWD band for 78% of the spring season, and the system achieved availability above 97.5% during 12 months of operation. The results show that locally produced sensing and open-platform control can improve irrigation-water productivity while remaining suitable for cooperative-scale deployment.

alternate wetting and drying, ESP32, Internet of Things, paddy rice, smart irrigation.