Expression of relationships between morphological traits is very important in scientific researches. The effective statistical method that can be employed in carrying out the task is "path analysis". In current study, at first 28 quantitative traits for 77 identified cultivars of pistachio in Kerman Province were selected and measured. According to importance, the trait of the number of pistachio in shell per one ounce was determined as dependent trait. By stepwise multiple regression, among 27 residual traits, 3 traits, including the cluster weight of fresh fruit, the number of pistachio in each cluster and the weight of cluster axis of fresh fruit which express the most variability of number of pistachio in shell per one ounce were selected (67.6 percentage). At first, path analysis was done on the basis of just-identified model (saturated) with four traits, i.e., the weight of cluster axis of fresh fruit, the cluster weight of fresh fruit, the number of pistachio in each cluster, and the number of pistachio in shell per one ounce. Then path coefficient related to direct effect of the weight of cluster axis of fresh fruit on the trait of the number of pistachio in shell per one ounce with view to the nature of traits and shortage of values related to it’s direct effect and total effect was considered zero. Thus, an over-identified model and testable model was created. Index value, GFI=0.98 confirmed goodness of fit of over-identified model. The results of the analysis of both models showed that these two traits i.e., the cluster weight of fresh fruit and the number of pistachio in each cluster have most direct effect on the value of the trait of number of pistachio in shell per one ounce, respectively.