Many of RNA functions depend on interactions between RNA and proteins. So, understanding the molecular mechanism of RNA-protein interactions (RPIs) is a maor challenge in structural bioinformatics. In this paper, we proposed a novel method for predicting RNA-protein interactions based on sequence information. e used motif information and repetitive site in RNA and protein sequences as features to build a model to RPI prediction using a random forest classifier. Results of 0-fold cross-validation experiments on two non-redundant benchmark datasets show the good performance of proposed method in RPI detection. Our method achieved an accuracy of and Matthews correlation coefficient (MCC) of 76.