The aim of this paper is to study the verb borrowing of Persian based on matrix language frame model. Myers-scotton (1993-2006) proposed matrix language frame model due to eminent role of one language in the bilinguals’ speech. Only one language as matrix language determine morpho-syntactic patterns of sentences containing the elements of two languages according to uniform structure principle of matrix language frame model. The data is corpus-based and they have been collected from Persian speakers’ speech in different contexts. The research method is descriptive-analytic. At first, Persian verb and its inflectional morphemes were classified into: content morphemes (non-verbal elements in compound verbs and bare form of the verb in simple verbs) and late outside system morpheme (inflectional morphemes of the verb and light verbs) according to differential access hypothesis or 4-M model. Then the data were analyzed based on the principles and hypothesis of matrix language frame model. Persian is matrix language in all data due to uniform structure principle of matrix language frame model. Persian acting as matrix language has caused the inflectional morphemes of its verb have been added to non-Persian verbs entered Persian verb construction through light verb construction and indirect insertion process. Matrix language has caused inflectional morphemes of Persian verb activated and selected at formulation level to assign Persian morpho-syntactic frame on the sentences. So non-Persian verbs have been activated and selected at conceptual level as content morphemes. These content morphemes are realized as infinite verbs in Persian speakers’ speech to carry meaning. It seems there is a relation between different morpheme nature (their selection at different abstract level) and their transferring to other languages. Myers-scotton (1993-2006) emphasizes this relationship. As he has predicted matrix language frame model is efficient in explaining language contact phenomenon so more expansive studies can be based on it.