The objective of present research is to conduct validity and reliability analysis of the verbal section of Test of Pretended Play that will measure pretended play behaviors of pre-school age children (3-6 years of age). Test of Pretended Play was first developed by Vicky Lewis and Jill Boucher in 1997. This test aimed to measure pretended play behavioral development of normal children between ages 1 to 6 besides children with mental developmental problems till the age 8. The test consists of two parts; verbal and nonverbal. Verbal part is composed of four subsections which are daily objects, toys and non-representative materials, one-symbol toy, playing pretended play alone. Reliability analysis demonstrated that test-retest correlation is 0.933 (p less than 0.001). In present research, at the end of statistical analyses conducted to detect the validity analysis of scale, a statistically meaningful positive relationship on level ,001 has been detected between Pretended Play Total Scores and Raven Progressive Matrix Test (RAVEN) total; a statistically meaningful positive relationship on level 0.001 with Ankara Developmental Screening Inventory (ADSI) and a statistically meaningful positive relationship on level, 001 with Language Use Scale. In order to compare with respect to pretended play normally developing children and children with autism and mental disability, abstract thinking, development and language use skills as the differences between total average scores are analyzed, it surfaces that score average of normally developing children is higher than children with autism and mental disability; score average of children with mental disability is on the other hand significantly higher than autistic group.