Clinical differentiation of facial lentigo senilis/initial seborrheic keratosis (LS/ISK), seborrheic keratosis (SK), lentigo maligna (LM), and lentigo maligna melanoma (LMM) can be difficult. Dermoscopy improves the diagnoses in pigmented skin lesions (PSLs), but it is not helpful for the sun-exposed face because of the flat rete ridges without network-derived features. Therefore, development of new diagnostic criteria for this particular localization is a current issue of dermatology. In this retrospective study, dermoscopic slides of facial pigmented skin lesions of 66 patients referred to two clinics in Turkey were evaluated. Our aim was to determine the reliability of dermoscopy in the differentiation of these entities. The facial PSLs of 66 patients (34 males and 32 females) (median age: 58.2) were photographed with a Dermaphot (Heine, Hersching, Germany) over a five year period from November of 1995 to May of 2000. All of the dermoscopic slides were analysed according to 27 dermoscopic criteria developed by Schiffner et al. This data set contained 22 histologically proven malignant (14 LM, 8 early LMM) and 44 benign (18 SK, 26 LS/ISK) PSLs. In general, asymmetric pigmented follicular openings, dark streaks, slate-gray streaks, dark globules, slate-gray globules, dark dots, dark rhomboidal structures, light brown rhomboidal structures, dark homogeneous areas and dark pseudonetworks were statistically significant for malignant growth. On the other hand, milia-like cysts, pseudofollicular openings, cerebriform structures, light brown globules, light brown dots, light brown homogeneous areas, yellow opaque homogeneous areas, and light brown pseudonetworks were statistically significant for benign growth. This research emphasizes that dermoscopic features on the face differ from criteria used in other locations of the body. Analysis of the data suggests that dermoscopy can be used in the differentiation of LS/ISK, SK, LM and LMM from each other.