Rapid and accurate estimation of rice Nitrogen Nutrition Index (NNI) is beneficial for management of nitrogen application in rice production. Traditional estimation methods required manual actual measurement data in the field, which was time-consuming and cost-expensive, and RGB images from unmanned aerial vehicle (UAV) provided an alternative option for nitrogen nutrition index (NNI) monitoring. In this study, RGB images from unmanned aerial vehicle (UAV) were obtained from each growth period of rice, and six machine learning (ML) algorithms, i.e., adaptive boosting (AB), artificial neural network (ANN), K-nearest neighbor (KNN), partial least squares (PLSR), random forest (RF) and support vector machine (SVM), were used to extract target information for estimating NNI as well as vegetation index (VI). Results showed that most UAV VIs were significantly correlated with rice NNI at the key growing periods; the estimation results of rice NNI using six ML algorithms showed that the RF algorithms performed the best at each growth period with the determination coefficient (R2 ) ranged from 0.88 to 0.96 and room mean square error (RMSE) ranged from 0.03 to 0.07, in which the estimation of NNI was the best in filling period and the early jointing stage. Rice NNI at the early jointing stage was significantly correlated with soil available nitrogen (AN) with the R2 of 0.84 in Pukou and 0.72 in Luhe, respectively, and rice NNI was significantly correlated with the yield with the R2 of more than 0.7 in Pukou at the whole period and more than 0.7 in Luhe from late jointing to maturity stage. Therefore, the combination of RGB images from UAV and ML algorithms was a scalable, simple and inexpensive method for rapid qualification of rice NNI, which effectively improved nitrogen use efficiency and provided guidance for precision fertilization in rice production.