This paper investigates a couple-group consensus problem of second-order multi-agent systems with the impact of second-order neighbours’ information. For systems with/without time delays, couple-group consensus criteria are established in the form of linear matrix inequalities by utilizing both model transformation and stability theories. The main results indicate that the group consensus for multi-agent systems under the effect of second-order neighbours’ information can be achieved based on the premise of a generalized balanced couple. Finally, illustrative examples are presented to demonstrate the effectiveness of the theoretical results.