Quality-related fault detection technology is an important mean to ensure safe operation and stable quality for industrial processes, which, thus, has recently become hotspots in the process industrial control domain. In this paper, an quality-related fault detection method based on adaptive mixed kernel canonical variable analysis(AMKCVA) is developed, considering the nonlinear and dynamic characteristics of industrial processes, as well as the varying characteristic of quality-related faults. Under that framework, mixed kernel functions and adaptive monitoring statistics are reasonably designed for improving the performance of quality-related fault detection. Finally, a case study on hot rolling process is given to demonstrate the advantages of the proposed approach compared with other methods.