A method for traveling time prediction is proposed using genetic network programming (GNP) based data mining. The method extracts the rules named Isolation Rules, that is, a kind of association rules having the consequent part with the narrow distribution of continuous values. A set of isolation rules is applied to continuous value prediction. The database of the traveling time of the focused route with traffic information is generated and isolation rules on the traveling time of the route are extracted. Traveling time prediction is done considering the matching rate of the isolation rules with the current traffic conditions.