The authors used longitudinal data to develop a risk factor index (RFI) for the prediction of smoking behavior in youth. Students were followed yearly from 6th through 12th grades in a prospective longitudinal study. Ten risk factors were identified and combined into an RFI. Data were analyzed with a generalized estimating equations approach. The RFI predicted both concurrent smoking and use of cigarettes in the succeeding year. It further predicted whether a youth would smoke at any point during his or her school career. Prediction was better for boys than for girls. Furthermore, the RFI better predicted heavy smoking than any use of cigarettes. The RFI could be useful in selecting youth for intensive prevention and early intervention efforts. Results also suggest the importance of further examination of gender differences in smoking behavior.