[Objective] To address the low efficiency of manual surveys and limited scale adaptability of existing remote sensing techniques in bund extraction, a multi-source data coordinated extraction framework of ‘macroidentification and micro-extraction' was constructed in order to enhance the accuracy and efficiency of farmland bund identification and parameter estimation. [Methods] At the county scale, satellite remote sensing imagery and digital elevation models（DEM） were used to develop a feature interpretation method for bunds（stone and soil bunds） and to analyze distribution response of bund types to terrain. At the small watershed scale, UAV-derived digital orthophoto maps（DOM） and digital surface models（DSM） were fused to extract geometric parameters of bunds, such as bund length, bund width, bund height, and bund coefficient. Additionally, the extraction accuracy was evaluated using deviation error（DE）, coefficient of determination（R²）, and root mean square error（RMSE）, and linear regression models between measured and inverted values of geometric parameters were constructed. [Results] ① Satellite remote sensing imagery demonstrated strong applicability for the identification and characterization of farmland bunds at the county scale. Validation in the Xiaziling small watershed indicated that recognition accuracies for both the number and area of bunds exceeded 91%. At Zhongxian County of Chongqing City, the total number of farmland bunds exceeded 300 000, with a total area of approximately 8 km² and a bund coefficient of about 3%.② UAV imagery enabled high-precision extraction of bund parameters at the watershed scale, with absolute deviations of less than 2 m for bund length, 0.1 m for bund width, 0.3 m for bund height, and 2% for the bund coefficient. The extraction accuracy ranked as follows: bund length > bund width > bund height > bund coefficient, with soil bunds exhibiting higher extraction accuracy than stone bunds across all parameters. The parameter regression models showed good fit（R² mostly above 0.78）, and RMSE was controlled within 1.6.③ At Zhongxian County, bunds were primarily distributed within elevation ranges of 300～600 m and slope ranges of 6°—15°, showing a trend of ‘first increasing and then decreasing' with increasing elevation and slope. Soil bunds had a wider distribution compared to stone bunds. Bund patches in valley areas had larger areas and higher numbers, while those in mountainous areas showed higher density but fragmented distribution. [Conclusion] The developed multi-source remote sensing coordinated method is suitable for the identification of bund distribution patterns at the county scale and for fine parameter extraction at the small watershed scale. The regression models can effectively estimate key bund parameters.