Mass transfer in osmotic dehydration of fruits at atmospheric pressure has been modeled either by a phenomenological approach using Crank's model (which is a Fick's law solution) or empirically, using models developed from polynomial equations, mass balances, or relations between process variables (i.e., Magee's, Azuara's, and Page's models). For all of these models, experimental data are required to determine the values of their adjustable parameters for specific processing conditions. From a wide variety of published data, the parameters for moisture loss (ML) and solid gain (SG) were calculated for the above-mentioned models, and a comparison of their correlation capability was made. Azuara's and Page's models yield better correlations (with mean absolute errors less than 1.26% for ML and 0.46% for SG) than Magee's and Crank's models (with mean absolute errors of up to 2.98 and 1.68% for ML and SG, respectively).