ABSTRACT Germination capacity is an important parameter of seed viability. The direct germination test destroys the seed. Viability evaluation by nondestructive tests allows the saving of seeds for storage and future use. However, predictive values of most nondestructive tests have not been adequately determined. This study investigated performance parameters and optimized assay conditions for a nondestructive colorimetric resazurin seed-viability test of flax seeds (Linum usitatissimum L.). In this study, 800 seeds were tested by resazurin test before germination. Optimization of resazurin test conditions and cutoff value was determined via a “receiver operating characteristic curve” analysis. We found that the optimal incubation time was 2 h. We proposed to measure sample absorption at 595 nm for better seed-viability estimation by resazurin test due to spectral properties of dye. We improved sensitivity of the resazurin test by using measurements adjusted to blank absorption at 595 nm (sensitivity 91%) or to sample absorption at 490 nm (sensitivity 95% versus 81% before corrections). The resazurin test reliably predicted germinable seeds in 72–73% cases and ungerminable seeds in 92–95% cases, with a sensitivity of 91–95% and a specificity of 71–74%. Therefore, the easy and low-cost resazurin test exhibited analytical performance similar to more expensive instrumental tests and can be suitable for nondestructive seed-viability estimation.