These codes include the entire process of soybean sample generation, including cloud computing GEE(Google Earth Engine) code and local computing python code. The GEE code includes code for generating random farmland points and non-farmland points, filtering potential soybean points and non-soybean points, and exporting their optical parameter time series curves as Geojson files, as well as feature calculation and feature classification. Python code was used to confirm generated soybean samples through parameter time series curve integration.