Sterilization of foods in cans has been well studied both experimentally and theoretically, but little work has been done on sterilization of food contain large particles in a fluid. Since many of canned foods belong to this group, we decided to focus on their sterilization. This work presents a global approach aiming at the characterization of the thermophysical properties and at the simulation of the coupled heat and mass transfers during thermal processing. The green ripe olive was selected as the model product in this study. At the first part of this study, the quality factors (firmness, color and antioxidant activity) the thermophysical properties of green olives (Conservalea cv.) were determined after thermal processing at 85, 90, 95 and 100°C for selected times (0, 10, 20 and 50) in a vertical cylindrical can ( 11.2 cm high and 9.8 cm in diameter). Penetration test was used for firmness measurement using Instron. Color (L, a and b) and anthioxidant activity of samples were evaluated with image processing method and with Sousa method (absorption of DPPH in 517nm), respectively. Thermal softening of olives follows from a first order kinetic reaction. Green (a) and yellow (b) color changes of olives surface follow from a first order reaction while their lightness (L) follows from a zero order reaction. The kinetic temperature dependence of these reactions was modeled by the Arrhenius equation. Thermal conductivity was modelled by the parallel model. At the second and main part of the study, the modelling of the coupled heat and mass transfers has been applied to a cylinder geometry using the finite elements method. Mathematical simulation is helpful in designing new or in improving existing processing systems or for their operation control. The two dimensional equations of mass, momentum and energy conservation are solved using FEMLAB software. The meshes used were 7882 triangular elements where 1056 of them were in boundaries. The fluid was assumed to have constant properties except for the viscosity (temperature dependent) and density (Boussinesq approximation). The predicted profiles of temperature were validated by comparison with the experiment. The comparison criteria used to evaluate goodness of fit, namely the coefficient of correlation ( r ) and the root mean square error ( RMSE ). The liquid flows upwards in a thin boundary layer and flows downwards in the interstice between the olives. The results showed that the natural convection means to be displaced the Slowest Heating Zone (SHZ) towards the bottom of the can, and eventually stay in a region that is about 10±15% of the can height from the bottom. The