Abstract: The soil water characteristic curve is one of the most crucial indicators in soil water dynamics and irrigation. However, the current measurement systems of soil water characteristic curves still had some challenges, such as the switching errors among multiple sensor probes, high costs, and limited fixed-points difficult in the field. In this study, an in-situ scanning system was designed to monitor the soil water at multiple depths in farmland using the frequency domain dielectric spectroscopy. The volumetric water content and matric potential data were collected from different depths. The scanning system also comprised a dielectric tube sensor module (including two sensors for the moisture and matric potential), a main control, a 4G communication, solar power, and an Alibaba Cloud monitoring module. The dielectric tubes were vertically installed in the soil to be tested. Two sensors were then utilized to scan and monitor the soil moisture and matric potential at various depths (5-60 cm, at 5 cm intervals). The data was collected to store locally on an SD card and then uploaded simultaneously into the cloud platform for remote monitoring. In system installation, the small-diameter drill bits (25 mm for the volumetric water content tube, and 85 mm for the matric potential tube) were used to bore holes for the dielectric tubes into the soil. A series of measurements were carried out in the various total depths, sensor probe intervals, and cycles. The low consumption of the system was obtained with the standby power at 0.62 W and working power at 2.4 W. Each operational cycle was about 2 min, with an hourly power consumption of 0.68 W·h. The capacity of the solar panel battery was 3×104 mAh with a rated voltage of 12 V. A long-term monitoring was realized in the field under adequate sunlight. The effective response radius of the sensors was determined to incrementally increase the soil layer height until the sensor output stabilized. The thickness of gypsum was 30 mm to measure the matric potential. The sensors were calibrated for the volumetric water content using soil samples with different water contents. A determination coefficient of 0.991 was achieved in the calibration curve. A commercial matric potential sensor (TEROS-21) was used to calibrate the soil matric potential. The determination coefficient of 0.989 was observed in the calibration curve, indicating excellent consistency within the measurement range. Field trials of the system were conducted in a demonstration field of returning green winter wheat. A total depth of 60 cm was measured with intervals and cycles of 5 cm and 1 h, respectively. The results indicate that the scanning system can accurately monitor the dynamic changes in the soil profile volumetric water content and matric potential within the root zone of winter wheat. The water characteristic curve of the soil profile showed that the water consumption in the 0-20 cm layer was attributed to both root absorption and surface soil evaporation. The main root absorption area of winter wheat in the regreening period was 25-45 cm, while the 50-60 cm layer with fewer roots maintained the higher soil water content. In-situ determination of soil profile moisture characteristic curves can be expected to assess the soil water retention capacity. The finding can also provide essential data and technical support for the soil moisture in the crop root zones during intelligent water-saving irrigation.