A control method for protecting failure recovery of Ethernet ring and Ethernet ring nodes are provided, which include: blocking ports recovered just now after detecting link has been recovered by failure link adjacent nodes, setting the ports as pre-forwarding state and starting a timer, unblocking the ports blocked temporarily and changing the ports from pre-forwarding state into forwarding state if the failure link adjacent nodes do not receive a control protocol message sent by a main node within the time that the timer timed. Application of the present invention can achieve that the ports of the link recovery adjacent nodes can come into forwarding state from pre-forwarding state as early as possible in the case of not receiving corresponding control protocol message in time, and avoid flow losing caused by that corresponding ports of failure can not forward data message in time effectively.
