Demand for soybean seeds with increased protein content is high in the international market. In this context, breeding soybean varieties targeting high protein content is the need of the day. In this paper, elite soybean germplasm accessions were screened for protein content. Eighty-eight soybean germplasm was evaluated for protein content through near infrared spectroscopy. A prediction model to determine the protein content of soybean germplasm through a non-destructive method was developed to calibrate the NIR. The protein content of 39 lines was analyzed in wet lab conditions and used to calibrate in NIRs between 1400 and 2400 nm at 2 nm intervals. High determination coefficient and low values of root mean square error (2.585) and standard error of prediction (2.832) confirmed the model’s utility for predicting the protein content of unknown samples. Accordingly, the protein content of 49 germplasm lines revealed that the genotypes LU96, TNAU20056, and SL525 depicted high values of 56.66, 55.51, and 54.48% protein content, respectively, but with fewer yields. The genotype RKS45 recorded a protein content of 45.33% with a single plant yield of 34.09 g, which can be further utilized for hybridization and selection. Thus, a non-significant correlation was observed between the protein content and single plant yield, suggesting that an increase in protein content will not directly influence the yield parameters. This paper provides a simple methodology to accurately determine the protein in a large set of samples in a short time, which helps in speed breeding programs.
