Objective: On the basis of Wilkins (2003) theory of dysfunction of visual pathway color perception, this quasi-experiment study attempts to assess the effect of the developed software on reduction of visual perception errors in students with dyslexia. Method: 40 students aging 6.7 to 9.10, from first, second and third elementary grades were selected and matched. Milaad Educational Software and Fernald traditional method were administered respectively for the experimental and control groups. Reading ability was evaluated by Reading Inventory (Pouretemad, 2001). Results: t-student analysis revealed that the experimental group showed an improvement in reading comprehension, speed and reduction of errors rather than the control group. Furthermore, the size effect estimation displayed that 22% of correct word reading variance, 21% of errors of reading variance, 27% of reading speed, and 14% of comprehension variance could be explained by the effect of Milaad software. Conclusion: Milaad newly developed educational software could be utilized for specific instructional purposes for students with dyslexia.