Objective: In this research the frequency of different cohesive devices in the written narratives of Farsi speaking students with and without learning disorders are compared. Method: The research method is Ex Post Facto or Causal Comparative. Halliday and Matthiessen's (2004) systemic functional approach was applied for analyzing the data. 25 boys with learning disorder and 25 boys without learning disorder were selected. The informants were 10 to 12 years old. Three researcher-made subtests (reading, visual, and listening) were administered for gathering the written data. Findings: Descriptive statistics were used to describe the data. Then Multi Analysis of Variation was used to analyze the significance between the differences of frequencies of cohesive devices. Results: Results showed that the difference of frequency of usage in the written narratives between students with and without learning disorder is significant if =α0/05.