The Forest Soil Moisture Monitoring Network (FSMMN) provides a coordinated, multi-depth dataset of soil moisture and temperature measurements from three eastern U.S. Forest Service experimental forests, including Hubbard Brook (New Hampshire), Fernow (West Virginia), and Coweeta (North Carolina). This latitudinal gradient spans distinct soil types, vegetation communities, and precipitation regimes across the Appalachian Mountains, capturing the variability in soil water dynamics. The network currently includes 44 monitoring sites and 262 soil moisture sensors distributed across the three forests, with hourly records beginning in 2022 at Coweeta and Fernow, and in 2023 at Hubbard Brook. At each forest, paired in situ volumetric water content (VWC) and soil matric potential (SMP) sensors were installed at three depth intervals (10–20 cm, 50 cm, and 60–100 cm) within multiple soil profiles across several catchments, enabling the characterization of soil moisture dynamics and hydraulic function at multiple scales. The data undergo automated and manual quality control procedures described in this paper and are updated annually in public data repositories. The dataset enhances the spatiotemporal coverage of soil moisture and temperature observations in forested headwater catchments where long-term, spatially distributed soil moisture records have historically been scarce. By capturing both vertical and lateral soil water variability across contrasting forest ecosystems, the FSMMN provides a foundation for cross-site studies linking soil hydraulic properties and catchment water balance at scales relevant to ecological and hydrological modeling.