The invention provides a cognitive radio ad hoc network node stability determining method based on a support vector machine. According to the cognitive radio ad hoc network node stability determining method, a stability state problem of the node in the cognitive radio ad hoc network is modeled based on the support vector machine according to local network attribute. The cognitive radio ad hoc network node stability determining method comprises an index parameter sensing and obtaining mechanism, and a node stability determining and routing deciding method based on the support vector machine. Firstly, through an interlayer interaction mechanism, the current network environment is sensed, and the index parameter is acquired and transmitted to a stability judgment decision module. Secondly, through a support vector machine model, the index is used as the input characteristic, and different stability states are used as classification labels for training, thereby obtaining a prediction output. Finally, different routing strategies are selected for performing data forwarding. According to the cognitive radio ad hoc network node stability determining method, through comprehensively analyzing the network environment, the node stability state is predicted according to a machine learning method, thereby improving adaptability of the cognitive radio ad hoc network and improving forwarding efficiency.