In the present study the performance of an air dehumidifier using lithium bromide (LiBr) as a desiccant was investigated. Response surface methodology (RSM) was used to assess individual and interactive effects of the six main factors (velocity, temperature and humidity of air, flow rate, temperature and concentration of desiccant) on dehumidification mass rate. A reduced quadratic statistical model was derived to predict dehumidification mass rate. The maximum dehumidification mass rate was obtained 0.154g/s under the optimal conditions of an air velocity of 4.1m/s, desiccant flow rate of 0.035kg/s, air humidity ratio of 0.0185kg/kg, desiccant concentration of 0.48kg/kg, air temperature of 29.5°C, and desiccant temperature of 21.8°C. The effectiveness number of transfer unit (NTU) model was employed to describe the coupled heat and mass transfer. The results of the model and the experimental data show good agreement. Dimensionless mass and heat transfer coefficients correlations are proposed; the average absolute differences between the predicted values and the experimental findings for Sh and Nu numbers were calculated as 2.14% and 5.27%, with the discrepancies mainly within ±9% and ±13%, respectively.