Experimental and modeling studies have been performed to determine mixing characteristics of binary mixtures in a spout-fluid bed. Spherical glass beads of diameters (3.075, 1.7, 1.2, and 0.75 mm) and air as fluidizing medium have been used in the study. Effect of various system parameters, namely, initial static bed height, gas velocity, diameter ratio, mixture composition, and sampling time on mixing of binary particles has been experimentally investigated. A dimensionless correlation has been developed for mixing index. Mixing behavior has been modeled using artificial neural networks (ANNs). Training of ANN was performed using the Levenberg-Marquardt (LM) backpropagation algorithm to predict the mixing index. The predictions of the ANN were found to be in good agreement with the experimental results and predictions from developed correlations.