Advanced controls have attracted increasing interests due to the high requirement on smart and energy-efficient (SEE) buildings and decarbonization in the building industry with optimal tradeoff strategies between energy consumption and thermal comfort of built environment. However, a state-of-the-art review is lacking on advanced controls for SEE buildings, especially considering advanced building energy systems, machine learning based advanced controls, and advanced occupant-centric controls (OCC). This study presents a comprehensive review on the latest advancement of advanced controls for SEE buildings, which covers recent research on data collection through smart metering and sensors, big data and building automation, energy digitization, and building energy simulation. Machine learning based advanced controls are comprehensively reviewed, including supervised, unsupervised and reinforcement learning, together with their roles and underlying mechanisms. In addition, advanced controls for energy security, reliability, robustness, flexibility, and resilience are further reviewed for energy-efficient and low-carbon buildings, with respect to fault detection and diagnosis, fire alarming and building energy safety, and climate change adaptation. Moreover, this study explores the advanced OCC systems and their applications in SEE buildings. Last but not the least, this study emphasizes the challenges and future prospects of the trade-off between complexity and predictive/control performance, AI-based controllers and climate change adaptation, OCC in thermal comfort and energy saving for the SEE buildings. This study offers valuable insights into the latest research progress concerning the underlying mechanisms, algorithms and applications of advanced controls for SEE buildings, paving the path for sustainable and low-carbon transition in building sectors.