Process mining has emerged as a key enabler of smart manufacturing. It provides techniques for process discovery, conformance checking, and performance analysis to enhance operational transparency and efficiency. With the growing availability of data from enterprise resource planning (ERP) systems, manufacturing execution systems (MES), and industrial Internet of Things (IIoT) platforms, manufacturers are increasingly leveraging process mining to enhance efficiency, quality, and flexibility. This paper presents a structured review of process mining in manufacturing, synthesizing one and a half decades of research. The reviewed studies are organized into core application areas, including discovery, conformance, performance analysis, and predictive or prescriptive analytics, with a particular focus on synergies with digital twins, operations research, simulation, and machine learning. We further identify domain-specific challenges, including heterogeneous data sources, high process variability, scalability, and real-time requirements, and discuss the limitations of existing approaches. Finally, we outline emerging directions—including online process mining, multi-level system integration, sustainability-driven analytics, and human–machine collaboration—highlighting how process mining can accelerate the transition toward smart and resilient manufacturing. This review consolidates prior work, reveals research gaps, and provides a roadmap for advancing process mining in the manufacturing domain.