Abstract: Rice plant height is an important indicator reflecting the growth status of rice. The current remote sensing retrieval methods of plant height are limited by data cost and update frequency, and can’t meet the requirements of large-scale and multi-frequency operational monitoring of crop plant height, and there is still a need to fully explore the potential of high spatial and temporal resolution optical satellite data. Choosing rice in the Northeast region as the research subject, measured plant height data from 2022 and 2023 were used, and an optical image-based feature set was constructed using Sentinel-2 data as the data source. Based on this, statistical analysis was first used to explore the sensitivity of each feature to plant height. Then, multiple regression models with different feature combinations were constructed using the random forest algorithm and the extreme gradient boosting algorithm. Simultaneously, feature recursive elimination and importance ranking methods were employed to analyze the importance of single features and different categories of features. Finally, the overall potential of Sentinel-2 images was evaluated. Finally, the random forest model with all features optimized achieved the highest estimation accuracy, with R²=0.857 and RMSE=8.395 cm. Comparative analysis revealed that among individual features, the vegetation water index and the blue band exhibited high importance. Red edge features, red band, green band, and texture mean were also selected in the optimal feature set, showing certain plant height monitoring capabilities. Among different categories of optical features, vegetation index features were more important than reflectance features. When texture features were applied alone, it was difficult to achieve high estimation accuracy, but they could serve as supplementary information when combined with the other two categories of features to improve estimation accuracy. Compared with previous studies, the accuracy of plant height estimation using Sentinel-2 data is comparable, but it enables rapid, multi-period, and large-scale estimation of plant height, showing good application potential.