Coordinated multi-point(CoMP) transmission is a promising technique to improve both cell average and cell edge throughput for long term evolution-advanced(LTE-A).For CoMP joint transmission(CoMP-JT) in heterogeneous scenario,if joint transmission(JT) users are firstly scheduled,other non-JT users will not be allocated sufficient resources,i.e.,scheduling relevancy exists in the users under different cells in the same coordination cluster.However,the CoMP system throughput will decline remarkably,if the impact of scheduling relevancy is not considered.To address this issue,this paper proposes a novel scheduling scheme for CoMP in heterogeneous scenario.The principles of the proposed scheme include two aspects.Firstly,this scheme gives priority to user fairness,based on an extended proportional fairness(PF) scheduling algorithm.Secondly,the throughput of the coordination cluster should be maintained at a high level.By taking the non-CoMP system as a baseline,the proposed scheme is evaluated by comparing to random PF(RPF) and orthogonal PF(OPF) scheme.System-level simulation results indicate that,the proposed scheme can achieve considerable performance gain in both cell average and cell edge throughput. Coordinated multi-point (CoMP) transmission is a promising technique to improve both cell average and cell edge throughput for long term evolution-advanced (LTE-A) .For CoMP joint transmission JT) users are previously scheduled, other non-JT users will not be allocated sufficient resources, ie, scheduling relevancy exists in the users under different cells in the same coordination cluster. Host, the CoMP system throughput will not remarkably, if the impact of scheduling relevancy is not considered.To address this issue, this paper proposes a novel scheduling scheme for CoMP in heterogeneous scenario. The principles of the proposed scheme include two aspects. Firstly, this scheme gives priority to user fairness, based on an extended proportional fairness (PF) scheduling algorithm. Secondarily, the throughput of the coordination cluster should be maintained at a high level. By taking the non-CoMP system as a baseline, the proposed scheme is evalua ted by comparing to random PF (RPF) and orthogonal PF (OPF) scheme. System-level simulation results indicate that that proposed scheme can achieve considerable performance gain in both cell average and cell edge throughput.