To solve the scheduling problem in single-arm cluster tools with chamber buffers and processing multiple part-type wafers, a branching search scheduling method was proposed based on the disjunctive graph model. First, a scheduling problem of cluster tools was transformed to a single machine scheduling problem. Next, a mathematical model of the problems was constructed based on disjunctive graphs. Then, a feasible solution space was obtained by using the branching method. After that, a branching searching algorithm was proposed for scheduling cluster tools with chamber buffers to obtain optimal robot action sequences to meet the system minimum makespan. Finally, experiments were designed to evaluate the proposed algorithm. The results indicate that the proposed algorithm is feasible and valid, and that the throughput of cluster tools with chamber buffers can be improved.