<p>The paper proposes a framework for the optimization of green wall plant selection and indoor air quality performance based on Gaussian Processes Regression and applicable in the context of Jordan’s Mediterranean climate for sustainable buildings. The research combines building simulation tools such as Energy Plus and CONTAM with machine learning modeling for the assessment of pollutant reduction, temperature stability, and humidity control in spaces with modular green walls. The results show that the inclusion of vegetation systems in buildings reduces critical pollutants such as CO₂ by 28%, VOCs by 33%, and PM₂.₅ by 39% compared to baseline values. High accuracy in the prediction results was obtained by the GPR model with R² and RMSE values of 0.94 and 0.12, respectively, which demonstrates its capability in modeling nonlinear interactions and uncertainty in the built environment. Moreover, the results show that plant species and air change rate (ACH) are the most significant factors affecting indoor air quality performance, which validates the capability of the AI-driven framework for plant selection optimization in the built environment. This research demonstrates the dual benefits of incorporating green walls in buildings in terms of environmental and health benefits and promotes the importance of considering the integration of vegetation systems in buildings for the development of sustainable buildings and spaces in support of the United Nations’ Sustainable Development Goals 11: Sustainable Cities and Communities.</p>