Abstract: Recurrent event data are often encountered in longitudinal follow-up studies related to biomedical science, econometrics, insurance claims, reliability and demography. The longitudinal pattern of gaps between successive recurrent events is often of great research interest. In this paper, we propose an alternative model which can accommodate ``cure fraction'' in gaps, and a nonparametric method is employed to the statistical inference. The proposed model and methodology are demonstrated by a small simulation result and the tumor recurrent data.