Aiming to address the issue of delayed detection of quality loss in traditional processing, it was desired to reduce losses by detecting defective products. However, it was found that losses had already occurred by the time defects were identified, making it difficult to achieve realtime quality control. The loss function model was proposed based on the quality characteristics of machining process. Taking the nitriding process of piston ring as an example, the main factors affecting the nitriding process were firstly analyzed by principal component analysis. These factors were then coupled with an asymmetric quality loss function to derive the quality characteristic loss model. Additionally, by considering the hidden quality loss within the tolerance range caused by processing errors, the hidden quality loss-production cost ratio was used to assess inherent reliability. The newly proposed loss model could predict the impact of processing on quality in advance and was expected to achieve realtime minimization of quality loss control objectives.