For the implementation of distributed cooperative control of the dc microgrid (DCMG), dependency on the communication layer is inevitable. Consequently, the DCMG, a typical cyber-physical system, is susceptible to cyber-attacks, which can hinder the achievement of overall control objectives, and the system may even get destabilized. To detect and mitigate cyber-attacks on actuators as well as sensors, a resilient mechanism is proposed in this article. First, the detection mechanism is proposed using a finite-state machine model-based technique that exploits a dynamic function. This function keeps investigating the actual and estimated values of actuator and sensor signals to authenticate the attack detection. Second, mitigation of attacks is carried out using a robust sliding-mode functional observer (SMFO). The signal of the compromised node is replaced by the SMFO estimated value in the event of an attack detection. The SMFO-based proposed technique is insensitive to measurement noise and parametric variations. The proposed approach is validated by simulation as well as experimental studies on a prototype of a 4-node DCMG.