In this paper, we study consensus problems of continuous-time multiagent systems. A multiagent system forms a communication network where each agent receives information from its neighbors. This information is used to obtain the control inputs necessary for achieving consensus. It is assumed that delays exist in communication networks. Communication delays are time dependent and differ for each communication channel. In addition, it is assumed that communication networks have switching topologies, and feedback gains are time varying. Under these assumptions, we show that a network system consisting of first-order agents is bounded and find a condition under which it achieves consensus. Stability is shown by using the Lyapunov theorem. In addition, we extend the consensus algorithm for first-order systems to an output consensus algorithm for high-order systems.