Canopy reflectance has been used in crops, such as corn and wheat, to assess crop status and direct in-season management practices, but less research has focused on using canopy reflectance in soybean research and production. In this study, soybean canopy reflectance measurements were measured at several growth stages throughout the 2015 and 2016 growing seasons using the RapidSCAN CS-45 Handheld Crop Scanner to determine if the normalized difference red edge (NDRE) index could be used to predict relative soybean productivity within a field prior to harvest. The NDRE values were used to calculate the cumulative reflectance of each experimental unit over the season. The cumulative reflectance readings through the R6 growth stage, termed the area under the reflectance progress curve (AURPC), and seed yield of every experimental unit were classified as top 25%, middle 50%, or bottom 25% within each location. Across all locations, bottom AURPC values correctly predicted bottom yield 52.5% of the time, and ranged from 46.7 to 86.2% by location. The probability of incorrectly predicting the bottom yield with a top AURPC value (9.7%) was also lower than incorrectly predicting the top yield with a bottom AURPC value (12.3%). Misclassifications by incorrectly identifying a bottom yield with a top AURPC ranged from 0.0% to 16.7% by location. Additionally, individual NDRE values at R2 were determined to be influenced by seed treatments at seven of the eight locations (p = 0.10) and, upon further investigation, found to be correlated to early-season soybean populations (r2 = 0.314).