This work aims to investigate the dynamics of water loss (WL) and solute gain (SG) during dehydration impregnation by immersion (D2I) and intermittent immersion (D3I). WL and SG of mango slices, 4 × 1× 1 cm3 in size, during dehydration immersion impregnation (D2I) and dehydration impregnation by intermittent immersion (D3I) were determined at 35, 45 and 55 °C for 270 min. Mango slices were immersed in a hypertonic sucrose solution of (61.6 ± 0.2) °Brix at a ratio of 6 mL of hypertonic solution per gram of fruit. Five semi-empirical models, two of which have been modified, were used to study mass transfer during D2I and D3I, namely the Azura, Weibull, Crank, Modified Crank I, and Modified Crank II models. The equilibrium water loss (WL∞), the equilibrium solute gain (SG∞), the diffusion coefficient, and activation energy were determined using the Modified Crank II model which was the best of all the tested models. Equilibrium water loss during D2I decreased with increasing temperature, while during D3I it increased with temperature. The SG∞ during D2I and D3I increases with temperature but was higher in D2I than in D3I. The average water diffusion coefficients were (6.02 ± 2.62) × 10−8 m2 s−1 for D2I and (4.89 ± 0.55) × 10−8 m2 s−1 for D3I and the average solute diffusion coefficients were (4.45 ± 0.53) × 10−8 m2 s−1 for D2I and (8.33 ± 0.79) × 10−8 m2 s−1 for D3I. The activation energies for WL in D2I and D3I were respectively 38789 J mol−1 and 9503 J mol−1. The respective values for SG were 9037 J mol−1 and 7327 J mol−1. This work demonstrates that mass transfers in D3I are better than those in D2I, and it highlights that, unlike the D2I process, the D3I process is less sensitive to temperature variations, making it particularly advantageous for processing products with high nutritional value.