For the problem that degradation state of the milling head cannot be directly measured， life data of milling head is difficult to obtain,sample size is small and cost is expensive,a model of milling head performance degradation and real-time life prediction based on the state-space model was proposed.The latest measured milling head temperature data were incorporated,the rate of change of milling head degradation and the failure probability density of the remaining life were updated using the Bayesian theory,then the online real-time life prediction of the milling head was realized.The temperature data during the operation of the functional milling head of the machining center shows that:the prediction accuracy of the remaining life increases gradually with the increase of the temperature data,the accuracy of the prediction is higher than 96% at the end stage.At the same time,the change rate of the degradation state gradually decreases with the increase of the running time of the milling head.