Data-driven and machine learning models have recently received increasing interest to resolve the bottleneck of computational speed faced by various physically-based simulations. A few studies have explored the application of these models to develop new, and fast, applications for fluvial and pluvial flood extent mapping, and flood susceptibility assessment. However, most studies have focused on model development for specific catchment areas, drainage networks or gauge stations. Hence, their results cannot be directly reused to other contexts unless extra data are available and the models are further trained. This study explores the generalizability of convolutional neural networks (CNNs) as flood prediction models. The study proposes a CNN-based model that can be reused in different catchment areas with different topography once the model is trained. The study investigates two options, patch- and resizing-based options, to process catchment areas of different sizes and different boundary shapes. The results showed that the CNN-based model generalizes well on “unseen” catchment areas with promising prediction accuracy and significantly less computational time when compared to physically-based models. The obtained results also suggest that the patch-based option is more effective than the resizing-based option in terms of prediction accuracy. In addition, all experiments have shown that the prediction of flow velocity is more accurate than water depth, suggesting that the water accumulation is more sensitive to global elevation information than flow velocity. Therefore, one can suggest that CNN-based models for flood prediction should consider large-size inputs and have large receptive field architecture to achieve a better performance.