Abstract A rapid and efficient detection method to detect the total starch in tobacco leaves using smartphone-based anthrone-sulfuric acid colorimetry is established under the framework of the low-cost, immediate, general-purpose, and high-throughput (LIGHt) smartphone-based colorimetry. The LIGHt smartphone colorimetry resulted average relative error of 5.74%, with a relative standard deviation (RSD) ranging from 2.58 to 4.31%, a detection limit of 1.53 µg/mL, and an average recovery rate of 95.72%. The results demonstrated that the LIGHt smartphone colorimetry perform comparably to the regular visible spectrophotometry in determining starch concentration, validating the feasibility of this novel colorimetric method, which requires no specific instruments. The LIGHt-based method indicated the feasibility and potential for application in the field of rapid quality detection of plant leaves and other food and medicinal-related quality control practices.