This study addresses the phylogenetic relationships among Pistacia species by amplified fragment length polymorphisms (AFLPs). A total of 31 wild Pistacia accessions belonging to P. eurycarpa, P. khinjuk, P. atlantica, P. mutica, P. integerrima, P. terebinthus, P. palaestina, P. mexicana, P. lentiscus species and to a hybrid between P. atlantica and P. integerrima and four P. vera cultivars were the plant materials of this study. Six AFLP primer combinations generated a total of 275 fragments, an average of 45.8 bands per primer pair, of which 254 (92.4) were polymorphic. Unweighted pair group method based on arithmetic average (UPGMA) and principle coordinates (PCo) analysis were performed using both mean character difference and Jaccard similarity matrices. According to the results, P. vera, P. khinjuk, P. eurycarpa, P. atlantica, P. mutica, P. integerrima and P. adantica x P. integerrima hybrids were in the first cluster. UPGMA analysis using mean character difference clustered P. palaestina, P. terebinthus, P. mexicana and P. lentiscus in the second cluster, whereas UPGMA analysis using Jaccard coefficient separated P. palaestina and P. terebinthus from P. lentiscus and P. mexicana. The P. khinjuk accessions had closer relationships to P. eurycarpa and P. atlantica than to P. vera which led to mis-identification of P. khinjuk samples as P. eurycarpa in this tudy. P. atlantica - P. mutica and P. terebinthus P. palaestina pairs were the closest species, and therefore P. mutica were classified as P. atlantica, and P. palaestina were as P. terebinthus. The resulted dendrograms and the PCo plots in this study did not support evergreen versus deciduous sectional division of Pistacia species and suggested classifying P. terebinthus in a separate group rather than being in the first cluster. Further study is inevitable including more evergreen species and accessions to clarify the position of P. terebinthus and the evergreen species.