Highlights What are the main findings? The Swin Transformer dual-branch deformable boundary network (STDBNet) proposed in this study improves the recognition performance for irregular terraces and small-scale targets, achieving significantly higher extraction accuracy (OA = 95.26%, MIoU = 86.84%) compared to mainstream semantic segmentation models. We constructed an annual spatiotemporal dataset of terraced fields on the Loess Plateau, covering nine time periods from 2017 to 2025, which enables a systematic analysis of their spatiotemporal variation characteristics. The terraces are predominantly distributed in low-altitude areas with gentle slopes, exhibiting a significant spatial coupling relationship with terrain gradient. What are the implications of the main findings? By integrating the Swin Transformer architecture, a dual-branch attention mechanism, and boundary-assisted supervision, the STDBNet model significantly enhances feature recognition accuracy for irregular terraces in complex terrain, offering a robust technical solution for terrace mapping and monitoring. The high-resolution annual terraced field time-series dataset, along with the spatiotemporal evolution characteristics it reveals, provides reliable data support and a scientific basis for studying soil and water conservation processes and optimizing ecological management on the Loess Plateau.Highlights What are the main findings? The Swin Transformer dual-branch deformable boundary network (STDBNet) proposed in this study improves the recognition performance for irregular terraces and small-scale targets, achieving significantly higher extraction accuracy (OA = 95.26%, MIoU = 86.84%) compared to mainstream semantic segmentation models. We constructed an annual spatiotemporal dataset of terraced fields on the Loess Plateau, covering nine time periods from 2017 to 2025, which enables a systematic analysis of their spatiotemporal variation characteristics. The terraces are predominantly distributed in low-altitude areas with gentle slopes, exhibiting a significant spatial coupling relationship with terrain gradient. What are the implications of the main findings? By integrating the Swin Transformer architecture, a dual-branch attention mechanism, and boundary-assisted supervision, the STDBNet model significantly enhances feature recognition accuracy for irregular terraces in complex terrain, offering a robust technical solution for terrace mapping and monitoring. The high-resolution annual terraced field time-series dataset, along with the spatiotemporal evolution characteristics it reveals, provides reliable data support and a scientific basis for studying soil and water conservation processes and optimizing ecological management on the Loess Plateau.Abstract Terrace construction is a critical engineering practice for soil and water conservation as well as sustainable agricultural development on the Loess Plateau (LP), China, where high-precision dynamic monitoring is essential for informed regional ecological governance. To address the challenges of inadequate extraction accuracy and poor model generalization in time-series terrace mapping amid complex terrain and spectral confounding, this study proposes a novel Swin Transformer-based Terrace Dual-Branch Deformable Boundary Network (STDBNet) that seamlessly integrates multi-source remote sensing (RS) data with deep learning (DL). The STDBNet model integrates the Swin Transformer architecture with a dual-branch attention mechanism and introduces a boundary-assisted supervision strategy, thereby significantly enhancing terrace boundary recognition, multi-source feature fusion, and model generalization capability. Leveraging Sentinel-2 multi-temporal optical imagery and terrain-derived features, we constructed the first 10-m-resolution spatiotemporal dataset of terrace distribution across the LP, encompassing nine annual periods from 2017 to 2025. Performance evaluations demonstrate that STDBNet achieved an overall accuracy (OA) of 95.26% and a mean intersection over union (MIoU) of 86.84%, outperforming mainstream semantic segmentation models including U-Net and DeepLabV3+ by a significant margin. Further analysis reveals the spatiotemporal evolution dynamics of terraces over the nine-year period and their distribution patterns across gradients of key terrain factors. This study not only provides robust data support for research on terraced ecosystem processes and assessments of soil and water conservation efficacy on the LP but also lays a scientific foundation for informing the formulation of regional ecological restoration and land management policies.