A Comprehensive Review of Computational Modeling, Data Analytics, and Intelligent Approaches to Greenhouse Ventilation Systems

• Review of 1,392 greenhouse ventilation studies published from 2016 to 2025. • 460 studies applied CFD, ML, or AI to analyze greenhouse ventilation. • Natural ventilation reduces energy use but depends on wind and geometry. • CFD simulations quantify air velocity and temperature gradients. • ANN and SVM models predict greenhouse microclimate and ventilation control. Ventilation is the primary mechanism to control temperature, relative humidity, and carbon dioxide levels in greenhouses. Properly designed ventilation systems remove sensible and latent heat generated by solar radiation and crop transpiration while maintaining adequate air exchange between indoor and outdoor environments. This review synthesizes research published between 2016 and 2025 on natural, mechanical, and hybrid ventilated greenhouses, with emphasis on the increasing use of computational and data-driven approaches. In this study, a comprehensive literature search was conducted across Web of Science, Scopus, and Google Scholar. A total of 1,392 publications on greenhouse ventilation were identified, including 460 studies focusing on computational fluid dynamics (CFD), machine learning (ML), or artificial intelligence (AI). The number of studies applying ML and AI increased after 2018. The review summarizes differences among ventilation strategies. Natural ventilation offers low energy demand and simple operation, yet its effectiveness depends on greenhouse geometry and weather conditions. Natural ventilation often causes spatial variability, affecting temperature distribution. Mechanical ventilation provides more consistent environmental control when natural ventilation cannot maintain desired conditions, particularly in regions with high solar radiation. Hybrid systems combine natural and mechanical ventilation, allowing flexible operation under changing environmental conditions. Computational fluid dynamics (CFD) has become an important analytical tool for examining airflow patterns, ventilation rates, and temperature distribution across greenhouse structures. However, CFD simulations are often run under steady-state conditions that do not fully reflect the transient conditions observed in commercial-scale greenhouses. In recent studies, machine learning models, such as artificial neural networks and support vector machines, were used to predict greenhouse microclimate variables from weather data and sensor measurements, enabling predictive control of vent openings and fan operation. Future research can focus on integrating validated computational models and real-time monitoring systems to support sustainable ventilation control across diverse greenhouse environments.