Abstract: Intelligent casting represents the core pathway for the transformation and upgrading of the casting industry under the background of Industry 4.0. Its essence lies in integrating technologies such as digital design, process simulation, online monitoring, artificial intelligence, and robotics into the entire casting production process to construct a data-driven closed-loop manufacturing system. Using a four-stage closed-loop framework of "design–process–execution–inspection" as the analytical structure, this paper systematically reviews the research progress of various technical routes in intelligent casting. In the digital design stage, additive manufacturing technologies represented by 3D sand printing have broken through the constraints of traditional molds on complex geometries, enabling rapid iteration from 3D models to sand mold forming. In the process simulation stage, multi-physics coupling modeling and digital twin technologies are driving the evolution of casting processes from offline simulation to online real-time monitoring. In the manufacturing execution stage, machine learning-driven melting control and collaborative robot applications are filling critical gaps in casting automation. In the quality inspection stage, deep learning methods have approached human-level performance in casting defect detection, yet data scarcity still restricts their large-scale application. On this basis, this paper summarizes the core challenges faced by intelligent casting across four dimensions: data, models, system integration, and industrial promotion, and provides an outlook on future research directions such as large-model-assisted process design and physics-data fusion modeling.