In the present study, an adsorption technique was explored to remove volatile organic compounds like toluene by using granular-activated carbon as an adsorbent. Experiments were conducted to draw breakthrough curves at varying bed height, inlet concentration, and flow rate to estimate the fraction of bed utilized till breakthrough. Response surface methodology technique was employed for optimizing the operating variables to maximize the bed utilization. Box-Behnken design methodology was used for the design of experiments. The higher value of regression coefficient (0.997) clearly demonstrated good agreement between experimental and model predicted response in addition to significance of the model. The optimum values came out to be: bed height 0.025 m; inlet toluene concentration 11,500 ppm; and gas flow rate 35 ml min(-1); respectively, at 77.4% utilization of bed till breakthrough. Moreover kinetic models such as Bed Depth Service Time, Thomas, and Yoon-Nelson were fitted into the experimental data. Their comparison revealed Yoon-Nelson model to be more appropriate for the prediction of breakthrough curves. Equilibrium data were found to fit the Langmuir isotherm satisfactorily.