Experimental and modeling studies have been conducted on CO2 absorption in a cyclone scrubber operated at room temperature. The effects of parameters such as the initial concentration of alkali in the solution and the liquid gas ratio on the CO2 absorbed flux were experimentally and theoretically investigated. A phenomenological model and three-layer feed-forward neural networks have been applied to estimate the CO2 absorbed flux in the cyclone scrubber. It was shown that the neural networks' values agreed well with the experimental data, while the values by phenomenological model partly agreed with the experimental data around the initial concentration of alkali in the solution, C(80)less than or equal to0.001 kmol/m(3) (pHless than or equal to11).