With the data source from highway charge system, the prediction model is established based on Bootstrap to improve the reliability of point prediction in travel time. Three indexes are used to evaluate the interval prediction performance, including prediction interval coverage probability(PICP), mean prediction interval width(MPIW), and coverage width-based criterion(CWC). Two key steps are analyzed and verified with actual data in modeling. As the methods used frequently in data-driven, the wavelet neural network and $ K $ nearest neighbor are compared about prediction error. The confidence interval prediction performance is analyzed among four kinds of Bootstrap methods. Under the same confidence level, the result shows that Percentile Bootstrap-KNN is the best with the minimum CWC. The proposed model is validated by Shanxi expressway in the case study. It is proved that interval prediction is better than point prediction under the same algorithm, as the reliability and value can be reflected by the prediction interval width.