This article aims to develop fully distributed adaptive protocols and establish the finite-time consensus criteria for nonlinear stochastic multiagent systems that are subjected to multiple uncertainties not only in the agents' states and inherent nonlinearities but also in their actuators. First, we propose a new continuous fully distributed consensus protocol for leaderless stochastic multiagent systems by leveraging the adaptive technology to enable the agents to achieve the finite-time consensus in the probability efficiently. Second, we develop an adaptive protocol for leader-following multiagent systems, and remarkably our protocol contains only one dynamic gain, which is adequate to resist multiple uncertainties and to ensure the realization of finite-time consensus in the probability. Finally, we provide numerical simulations to verify the effectiveness of our protocols.