Study on multi-agent systems is considered by researchers in the past decades . These systems were initially only used in military applications , but civilian applications such as air traffic control , distributed sensor networks , mobile robot networks and unmanned aerial vehicles have been also considered in recent years . The problem of controlling multi-agent systems , is as important as any other system . Distributed cooperative control is one of the important control methods in these systems , which control algorithm for each agent is obtained based on the information of its neighboring agents and the communication-graph topology between them . Consensus problem is one of the important problems in multi-agent systems . Consensus is defined as to reach an agreement on a specific quantity that depends on all the states of the agents . Also some faults may occur for a multi-agent system . If the fault is not identified , resolved or compensated , it may cause system instability and degradation . Today , design of distributed fault-tolerant control (FTC) based on the information of neighbors of each agent is known as an effective control method to compensate for the faults of individual agents , in addition to satisfying the consensus . By using adaptive control as an FTC method , the actuator faults can be compensated online and without the need of fault identification . Also it is not necessary to know the bound of external disturbances , uncertain dynamics and unknown nonlinear functions . According to these points , in this research , solving the consensus problem of nonlinear multi-agent systems with nonlinear functions , external disturbances and heterogeneous dynamics of followers is considered . For this purpose a distributed neuro-adaptive fault-tolerant controller is designed to compensate actuator faults . Key Words: Multi-Agent Systems, Consensus, Fault-Tolerant Control, Actuator Fault, Adaptive Control