Utilization of geosynthetics as a reinforcement material to improve the weak subgrades is gaining popularity over decades. The evaluation of the improvement achieved from these foreign materials is always a difficult task. It is also necessary to have a quantitative assessment of the improvement achieved from these techniques to adopt them in the new designs. This paper presents the results of two prediction models such as artificial neural networks (ANNs) and non-linear multiple regression models to estimate the bearing capacity of a foundation resting on geocell reinforced sand beds. Modeling results are compared with the experimental results to address the efficacy of these models. Overall, results obtained from both models showed good agreement with the experimental measurements up to about 10% of the footing settlements. At higher settlements, ANNs prediction is better than the multiple regression models.