Kiwifruit ( Actinidia deliciosa ) is a fruit full of various vitamins. Due to the short shelf life and fast deterioration, its storage and processing before consumption is of great importance. Osmotic dehydration process is one of preservation techniques for fruits, which is mostly used as a pretreatment for subsequent processes such as drying. In this process, both water loss and solid gain phenomena occur simultaneously. However, the low rate of water loss in the process is one its drawbacks. Applying pretreatments such as ultrasound waves can help overcoming this problem. In order to reach the optimum performance of each process, the control of the process is needed which in turn requires investigating on the mechanisms involved in the process in detail and knowing how the effective parameters act on it. In this regard, adopting a phenomenological modeling approach can facilitate the study. Moisture sorption isotherms are important tools to determine changes in moisture content during storage of the food materials, thermodynamic conditions after mixing the products and making decisions pertaining to undesirable chemical and enzymatic reactions. On the other hand, determination of the glass transition temperature ( T g ); as a powerful tool for understanding water properties in foods and control of the products’ shelf life, is vitally important. The objective of the present study was applying osmotic dehydration process for kiwifruit slices in sucrose solution at concentrations of 30, 40 and 50 °Brix, immersion durations of 0, 5, 10, 20, 30, 60 and 120 min, and indirect ultrasound pretreatment at durations of 10, 20 and 30 min. After conducting the experiments, the optimum treatment was determined using response surface methods (RSM) and genetic algorithms (GA). Through analytical modeling of osmotic dehydration process in axisymmetric cylindrical coordinates (two-dimensional approximation), the coefficients of moisture diffusion were obtained. The effect of osmotic dehydration and ultrasound pretreatment on fruit’s tissue was examined by scanning electron microscopy (SEM). Moisture desorption isotherms for the fresh and optimized treated samples were determined at temperature of 20 °C. Among well-known models, the most appropriate one was selected based on the highest correlation coefficient and the lowest root mean square error and mean relative error indices. The results showed that applying the ultrasound waves before osmotic dehydration of the product enhanced the rate of water loss. Moreover, by increasing the immersion duration, concentration of the osmotic solution and ultrasound pretreatment time water loss and solid gain increased. The best model predicting the moisture desorption of the kiwifruit samples were GAB and Peleg models. Thermal analysis showed that the T g values decreased with increasing moisture content. The modeling of T g showed that the best model for predicting the variations in T g values of the fresh and optimized treated samples with equilibrium moisture content was Brostow et al model. The obtained values of model parameters Gordon - Taylor near other researchers parameters better. Moreover, in modeling of T g as a function of water activity, Ross model and Khalloufi et al. model showed an acceptable agreement with the experimental data. Key Words: Glass transition temperature, Moisture desorptionption isotherm, Optimization, Thermal analysis.