Fm-transform approximation loses its accuracy near the boundaries of the function domain. This is because of the truncated basic functions in the first and last subintervals. In this paper, we impose some effective techniques on Fm-transform to overcome this defect. Throughout various examples, it is shown that the utilized techniques can significantly improve the accuracy of the Fm-transform approximation on the boundary subintervals.