Bubble size is one of the key parameters in the design of two-phase gas-liquid bubble column reactors. Accurate knowledge of this parameter is essential for the prediction of gas holdup, heat and mass transfer coefficients. The previous findings, particularly with respect to the influence of orifice size and physical proper-ties of the liquid phase on bubble size, are often of contradictory nature. In this paper, extensive new experimental results are presented for regions where published data are insufficient. The suitability of artificial neural networks for identification of the process variables and modeling is evaluated. The Radial Basis Function (RBF) neural network architecture was used successfully to generate a nonlinear correlation for the prediction of bubble diameter. This correlation predicts the present data and the control data of other investigators with excellent accuracy.