Abstract: Plasma spraying technology has gained widespread application in industrial fields due to its high efficiency and versatility. By utilizing ultra-high temperatures and high-speed jets, this technique rapidly heats and accelerates metal or ceramic particles to form high-performance coatings. However, optimizing coating quality remains a complex challenge, as its microstructure is significantly influenced by particle flight behavior, spraying parameters, and substrate conditions. This study aims to provide theoretical insights into the relationship between particle flight dynamics and coating performance, facilitating the precise control of coating microstructures. This review summarizes the acceleration, heating, spreading, and solidification behaviors of particles during flight and upon substrate impact. Employing particle diagnostic techniques and multiphysics simulations, it quantitatively analyzes the effects of spraying parameters-such as torch current, gas flow rate, and spraying distance—on particle velocity, temperature, and morphology. Additionally, the intrinsic relationships between particle flight characteristics (e.g., Reynolds and Weber numbers), spreading behavior, and interfacial heat transfer mechanisms regulating coating microstructures are thoroughly discussed. The application of machine learning and deep transfer learning techniques in plasma spraying is also highlighted. Advancements demonstrate that precise control of particle flight temperature and velocity is crucial for enhancing coating bonding strength and density. Moreover, the thoughtful design of spraying parameters can significantly reduce defects such as porosity and microcracks. This review not only deepens the understanding of particle flight behavior and coating formation mechanisms but also offers theoretical guidance and technical strategies for the development of advanced ceramic coating systems. Future studies should focus on deconstructing the relationship between particle melting states, splashing behavior, and coating performance. High-precision simulations of droplet solidification in dynamic environments could uncover additional mechanisms. Future research should fully consider real-world experimental conditions in simulations and establish more realistic multiphysics models. By utilizing high-time-resolution experimental techniques, dynamic data can be used to calibrate and optimize simulations. Expanding experimental parameters to include more spraying conditions will strengthen simulation validation. Real-time monitoring technologies can be used to feed experimental data directly into models, facilitating the development of adaptive, real-time simulation tools. Systematic sensitivity analyses should identify the main factors driving discrepancies between experimental and simulation results, and a comprehensive workflow from experimentation to simulation and process optimization should be developed to provide clear industrial guidelines.