Steel is widely used as building material for large-scale structures, such as oil and gas pipelines, due to its high strength-to-weight ratio. However, corrosion attack has been long recognized as one of the major reasons of steel pipeline degradation and brings great threat to safety in normal operation of structure. To mitigate the corrosion attacks, coatings are generally applied to protect steel pipelines against corrosion and improve durability of the associated structures for longer service life. Although have higher corrosion resistance, coated pipelines will still get corroded in a long run, as coatings may subject to damages such as cracks. Cracks on coatings could lower the effectiveness of protection for associated structures. Timely updates of up-to-date corrosion rate, corrosion location, and coating conditions to the pipeline risk management model and prompt repairs on these damaged coatings would significantly improve the reliability of protected structures against deterioration and failure. In this study, a corrosion risk analysis system is developed to detect and locate the corrosion induced coating cracks on coated steel using embedded fiber Bragg grating (FBG) sensors. The coatings investigated include high velocity oxygen fuel (HVOF) thermal sprayed Al-Bronze coating, wire arc sprayed Al-Zn coating, and soft coating. Theoretical models of corrosion risk assessment system were carried out followed by systematic laboratory experiments, which shows that the developed system can quantitatively detect corrosion rate, corrosion propagations, and accurately locate the cracks initialized in the coating in real time. This real-time corrosion information can be integrated into pipeline risk management model to optimize the corrosion related risk analysis for resource allocation. To place the sensing units of the system in the most needed locations along the huge pipeline systems for an effective corrosion risk assessment, an example case study is conducted in this study to show how to locate the most critical sensor placement locations along the pipeline using worst case oil and gas discharge analysis. Further applications of the developed system can be integrated with pipeline management system for better maintenance resource allocations.