Abstract: Markov models of Panel-data were proposed to resolve performance forecasting on asphalt pavements. The data was detected from north and middle region of Liaoning province. Intercept and coefficient of the disintegration function were calculated by a particular software named E-View which was based on the theory of panel-data. Markov state-transition matrix of calculating performance forecasting of asphalt pavement was conducted by MATLAB, in which the probability of the status of a specific road segment was predicted. The result shows that deterioration laws of each road segments in a same subset are similar. The problem of lack of data of accumulative years can be solved by utilizing the increasement of the sample data in a same subset. In order to test the prediction accuracy, the forecast data of the test samples were compared with the detection data of the same samples which were detected in 2014. The trial sample error was within ±5, which means the performance forecasting of asphalt pavement can be predicted by this method when the data of accumulative years is insufficient.