The photoelastic determination of S.I.F. in Fracture mechanics has been regarded as one of the most effective and practical experimental methods in which stresses are read directly, except a few shortcomings involved in the process of experiment; the difficulties of making a sharp crack tip similar to the practical one and nearly impossibilities of carving an arbitrarily shaped crack on the test plate, etc. To eliminate flaws mentioned above, recently, Kitagawa and Watanabe of Tokyo Univ.developed a new method named"Teflon Insert Method". which has improved experimental accuracy to a considerable extent byt remaining still room for further improvement, that is, the elimination of bonding boundary scars which render photoelastic fringes obscure. In this paper, a newly exploited"Teflon Molding Method" was attempted for the completion of teflon-epoxy experimental method. The experimental results obtained by this method are compared with existent theoretical and experimental values to evaluate its accuracy. As result, 1-6% of margin of errors were appeared in a series of photoelastic experiments which defied any other conventional method in terms of experimental accuracy.perimental accuracy.