Process alarm systems and process monitoring play a significant role in ensuring stability and safety in industrial production. The continual increase in the complexity and scale of industrial processes makes it difficult to monitor processes online and accurately track the root causes of alarms. To address these problems, online distributed process monitoring and alarm analysis using novel canonical variate analysis (CVA) with multicorrelation blocks (MCB-CVA) and enhanced contribution plot is developed in this paper. The proposed methodology is simple: first, process knowledge is used to divide the entire system into multicorrelation blocks to enhance fault detection performance; an improved contribution plot method based on principal component analysis is adopted as an alarm analysis strategy such that the distributed contribution can be used for online analysis; finally, a model of distributed alarm analysis is developed using the novel MCB-CVA and enhanced contribution plot. The performance of the proposed methodology is tested using case studies based on the Tennessee Eastman process (TEP). The simulation results show that the proposed MCB-CVA can achieve good online performance in not only fault detection but also root cause analysis, proving the practicability and effectiveness of the proposed MCB-CVA model.