Nitrogen nutrient index (NNI) is widely used as a good indicator to evaluate the N status of crops in precision farming. However, interannual variation in weather may affect vegetation indices from sensors used to estimate NNI and reduce the accuracy of N diagnostic models. Machine learning has been applied to precision N management with unique advantages in various variables analysis and processing. The objective of this study is to improve the N status diagnostic model for winter wheat by combining remote sensing data with weather data using random forest regression. N plot experiments including five nitrogen levels (0, 120, 180, 240 and 300 kg N ha^-1) and four winter wheat varieties were conducted from 2016 to 2018 in Laoling County, Shandong Province. North China Plain in Laoling County, Shandong Province, China. The selected vegetation index in this study was normalized difference vegetation index (NDVI), normalized difference red edge (NDRE) and the corresponding nitrogen sufficiency index (NSI). The results indicated that incorporating weather data improved the performance for NNI estimation using random forest regression (R²=0.82-0.85), compared to only using NDVI or NDRE (R²=0.53-0.55), The Random Forest Regression Model based on NSI calculated with NDVI and NDRE and weather data obtained the best nitrogen diagnostic performance with area agreement (83%) and kappa coefficient (0.677) all N rates, varieties, stages, and years. This study is to provide a basis for precision N management and help the green and sustainable development of agriculture. More studies are needed further to evaluate it under diverse on- farm and soil and weather conditions.