We present a new noise-resilient secondary control scheme for voltage and frequency synchronisation of an autonomous microgrid (MG). The communication network is an integral part of distributed secondary control structure. The communication links among the distributed generator units are assumed to be ideal in nature. However, the communication links are subjected to uncertain noises, which can significantly affect the synchronisation performance of MG control. We consider the information received over the communication link is corrupted with generalised Gaussian noise. Further, the complete non-linear model of the MG system has been considered for designing a robust distributed control scheme augmented with an auxiliary corrective control input to counterbalance the impact of noises. The performance of the proposed control scheme is evaluated by pursuing simulation of a MG test system in MATLAB/SimPoweSystem environment under load perturbation, changes in communication topology, communication link delays, data dropout, and variations in noise parameter. The results of a comparative assessment of the proposed control scheme with two approaches first a neighbourhood tracking error-based control scheme and second a noise-resilient control scheme are presented to illustrate the effectiveness of the proposed robust noise-resilient control scheme in voltage and frequency synchronisation and accurate active power-sharing. © The Institution of Engineering and Technology 2018.