In Northeast Thailand, approximately 90% of paddies are rainfed, and planting activities depend on the onset of the rainy season. As a result, planting dates exhibit considerable spatial and temporal variability. However, previous studies estimating rice production in this region have rarely accounted for such variability, mainly due to the lack of observational data with sufficient spatiotemporal distribution. To address this limitation, we used PhenoRice, a satellite-based model that estimates rice planting dates by analyzing MODIS time series. PhenoRice identifies phenological stages using the Enhanced Vegetation Index (EVI), which reflects canopy greenness, and the Normalized Difference Water Index (NDWI), which indicates surface water. Nevertheless, in Northeast Thailand, most paddies are established using dry direct seeding (DDS), where no standing water is present before planting. This contrasts with the assumptions of the original PhenoRice model, which relies on detecting pre-planting flooding. Furthermore, PhenoRice uses 250-meter MODIS data, which makes large-scale analysis over the entire 170,000 km² region of Northeast Thailand impractical. Therefore, we developed a modified version of PhenoRice that is calibrated with local observations, adapted for DDS systems, and capable of using 1-kilometer MODIS data for wide-area application. The model was validated using province-level monthly planted area statistics provided by the Office of Agricultural Economics, Thailand. The results showed a correlation coefficient of 0.59 and a root mean square error (RMSE) of 631 km², equivalent to about 10% of the regional rice area. However, model performance declined in 2002 and 2019, likely due to drought and typhoon-induced flooding. This indicates that while the model is effective under normal conditions, its accuracy is limited under extreme weather events during the growing season. The estimated planting dates from this model could be integrated into crop simulation models such as ORYZA2000 to improve rice yield estimation based on realistic planting conditions.