Soil organic matter (SOM) is crucial to soil quality. Traditional studies on SOM spatial distribution prediction establish models using the mean values of remote sensing images collected over fixed periods, which neglects the influence of seasonal variations and vegetation dynamics to a certain degree. Therefore, selecting an optimal time window is important for SOM spatial distribution prediction research. The North China Plain suffers from high uncertainty in remote sensing data due to its short bare soil period and frequent cloud and haze occlusion, and relevant studies on this region are still limited. Taking Yucheng City in Shandong Province as the study area, this research classified time windows for remote sensing images based on vegetation and cultivated land conditions. Using 2017 Landsat-8 imagery and the random forest model, we evaluated the accuracy and differences of SOM mapping based on remote sensing data from different time windows. Finally, extreme climate factors were introduced to explore their impacts on the spatial distribution of SOM. The results indicated that when only remote sensing variables were used, the SOM prediction accuracy of different time windows ranked as: reviving and sowing period > vigorous growth period > fallow and maintenance period > maturity and harvesting period, with the coefficient of determination (R²) ranging from 0.257 to 0.330. After incorporating environmental covariates, the prediction accuracy of all time windows increased significantly, with R²ranging from 0.413 to 0.477. The reviving and sowing period achieved the highest accuracy, proving that February to May is the optimal time window. When extreme climate variables were further added on the basis of the optimal time window, the R²of SOM prediction increased to 0.501. The number of warm nights and monthly minimum daily minimum temperature were identified as key variables affecting the spatial distribution of SOM. Focusing on the unique environmental conditions of the North China Plain, this study elucidates the effects of time windows and extreme climate factors, and offers new insights and methodsfor research in similar regions.