This paper proposes a novel control scheme for frequency support in MTDC grid-based hybrid ac/dc networks. This control is based on the multi-agent system (MAS) paradigm with an adaptive frequency droop control, and is distributed in all grid-tied converters. The objective of this control is to enable the ac grids to intelligently decide on their participation on frequency support, considering their power reserves, demand, and relevant technical constraints and requirements. The main achievement of the proposed control is to minimize the rate of change of frequency (RoCoF) and frequency undershoot (f nadir), and also providing a systematic enhancement in frequency stability, particularly in the disturbed and weak ac grids. A stability analysis is performed for the proposed control, using vector Lyapunov method for interconnected systems, deriving the conditions for stable operation. Test scenarios are conducted on a modified 7-terminal C IGRE g rid, c onnecting fi ve ac gr ids. Re sults un der di fferent scenarios show the performance and robustness of the proposed control in providing significant enhancement in frequency regulation for the connected ac grids. Nomenclature Symbol Description f nom,i Nominal frequency of ac grid i ∆f i Frequency deviation from f nom,i ∆f max,i Maximum instantaneous ∆f i P gen,i Synchronous generation in ac grid i P mtdc,i Wind power delivered to ac grid i P dem,i Power demand in ac grid i P crit,i Critical power demand for ac grid i f H (P dem,i) Probability density function of power demand in ac grid i at hour of day H a i Agent representing ac grid i G Graph for multi-agent ecosystem N i Set of immediate neighbors of a i R i Set of requesting agents to a i V Set of agents in MTDC grid P req,ij Power request from a i to a j P supp,ji Power support from a j to a i (as a reply to the request P req,ij