In this work, a two-step chemical-potential-gradient model based on nonequilibrium thermodynamic principles was developed to investigate the dissolution mechanism of crystalline active pharmaceutical ingredients (APIs). The perturbed-chain statistical associating fluid theory was used to calculate the required solubilities and chemical potentials of the investigated APIs. The statistical rate theory was used to describe the mass-transfer rate of the APIs at the solidliquid interface during the dissolution process. Dissolution profiles of indomethacin, naproxen, and glibenclamide in water and in buffered solutions at pH 5.0, 6.5, and 7.2 were measured using a rotating-disk system (USP II). The specific dissolution mechanisms of the APIs, such as surface reaction and diffusion, were analyzed by applying the proposed model to identify the rate-controlling step. The results show that the dissolution mechanisms of indomethacin, naproxen, and glibenclamide change with varying pH values of the solution medium. On the basis of the calculated rate constants, the dissolution profiles were modeled with a high degree of accuracy when compared with the experimental data.