Active crop canopy sensors have been used to non-destructively estimate nitrogen (N) nutrition index (NNI) for in-season site-specific N management. However, it is time-consuming and challenging to carry the hand-held active crop sensors and walk across large paddy fields. Unmanned aerial vehicle (UAV)-based remote sensing is a promising approach to overcoming the limitations of proximal sensing. The objective of this study was to combine unmanned aerial vehicle (UAV)-based remote sensing system and Crop Circle ACS-430 to estimate rice (Oryza sativa. L.) N status for guiding topdressing N application in Northeast China. Two N rate experiments involving two different varieties were conducted in 2014 at Jiansanjiang Experiment Station of China Agricultural University, Heilongjiang Province, Northeast China. An active canopy sensor Crop Circle ACS-430 with three spectral bands (red(R), red edge (RE) and near infrared (NIR)) and an Octocopter UAV equipped with a Mini Multi-Camera Array (Mini-MCA) imaging system with five spectral bands (blue (B), green (G), R, RE and NIR) were used to collect reflectance data at the panicle initiation (PI) and stem elongation (SE) stages. The preliminary results indicated that Crop Circle ACS430-based vegetation indices (VIs) explained 79-80% and 86-87% variability of aboveground biomass (AGB) and plant N uptake (PNU), respectively, but had very poor relationship with plant N concentration (PNC) (R² = 0.16-0.21) across all stages. The N sufficiency index (NSI) calculated with Crop Circle ACS-430 vegetation indices (NNI-VIs) had better correlation with NNI than the original VIs, especially at SE stage and across both stages, with the best R² of 0.65 and 0.69. UAV-based remote sensing VIs could be used to estimate Crop Circle VIs and NSI-VIs very well at both growth stages. The NSI_VIs-NNI approach performed well for diagnosing rice N status. Combining UAV-based remote sensing system and Crop Circle ACS-430 had a good potential for in-season diagnosis of rice N status at PI stage, with the highest accuracy rate (90%) and kappa statistics (0.62), but did not perform well at SE stage and across both stages. More studies are needed to further evaluate these different strategies.