Abstract: Process mining is to extract structured process information from the actual business event log. Process mining technology is now widely used in the process discovery and assisted modeling of real business processes，and can help to improve existing business processes through differential analysis. This paper proposed a cluster-based process mining method. Firstly，it described the activity-based process pathway starting from the event log，then analyzed the pathway clustering of the active view awareness based on distance，and analyzed the characteristics of process through the clustering results. These can provide decision making for process change. This paper took the actual conference registration system as the experimental object，demonstrated the effectiveness of the proposed method，it provides support for decision making for process modification.

Key words: business process, clustering, elbow method, conformance checking