Background: Intrinsically disordered proteins (IDPs) that do not possess stable secondary and tertiary structures are crucial for the function of numerous proteins.Even though these proteins lack intrinsic structure, they could bind to many different macromolecular partners when functioning in protein and protein interactions.And their prevalence is implicated in a number of human diseases.Accurate prediction of disordered regions from protein sequence is a necessary prerequisite for the further understanding of principles and mechanisms of protein function, and is a key for the elaboration of a structural and functional hierarchy of proteins.Therefore prediction of IDPs has attracted the attention of many researchers, and a number of prediction methods have been developed.Predictions of disorder play major roles in directing laboratory experiments that are leading to the discovery of disordered proteins, and thereby leading to a positive feedback loop in the investigation of these proteins.Methods: In this work, we propose a novel predictor, DRPred, which is based on a disorder position-specific scoring matrix (DPSSM).We develop a strategy to construct DPSSM which is generated by sequence alignment, and its element is disorder profiles.DRPred uses a database of disorder states of 35811 entries that is a non-redundant database contains amino acids and relative one of disorder, near-disorder and order information.A query is aligned against the database and resulting in DPSSM.The training set is compiled by Pierre and colleagues.It contains 723 protein sequences with 215,612 residues, of which 13,909(6.5%) are disordered.Our approach utilizes a custom-designed set of features that are based on sequence profiles, predicted secondary structure, a structural position-specific scoring matrix (SPSSM), DPSSM, predicted shape string and its profiles.Results: The PDB723 set is used as the training in the classification algorithm of Conditional Random Fields (CRFs).Short and long disordered regions are predicted by using our unitary predictor.The performance of our approach for predicting disordered regions exhibited Matthews correlation coefficient (MCC) of 72.55%, sensitivity (SN) of 70.63%, specificity (SP) of 98.63% and accuracy (AC) of 96.82%.Conclusions: DRPred outperformed on accuracies of prediction when compared with well-established publicly available disordered region predictors.It can be especially useful for prediction of disordered regions .