In this study, we propose a novel approach, namely the combined Convolutional Deep Autoencoder–Echo State Network (CDAE-ESN) model, for the analysis and forecasting of dynamics and low-order statistics in coupled turbulent channel-porous media flows. Such systems find wide applications in industrial settings, including transpiration cooling and smart interface engineering. However, the complex geometry of coupled flow systems presents additional challenges for purely data-driven models. Our results demonstrate that the integration of deep autoencoder and echo state network techniques enables effective modeling and prediction of dominant flow behaviors, particularly within the porous domain exhibiting laminar regimes. To enhance the model’s applicability across a broader range of data domains, we further employ fine-tuning on a dataset encompassing varying porosities. The achieved average statistics exhibit a reasonable agreement, underscoring the efficacy of our proposed approach.