For the past few decades, self-consistent field theory (SCFT) has been a popular tool for the study of polymeric nanostructures. However, SCFT is intrinsically a mean-fields theory, and there are limitations in applying it for complex nanostructures which are strongly affected by compositional fluctuation. Recently, a few methods have been developed to include fluctuation effects on polymer field theory, and Langevin field theoretic simulation (L-FTS) is one of the front runners. By adopting partial saddle point approximation and Langevin dynamics on polymer field theory, L-FTS has successfully incorporated the fluctuation effect. Even though it is faster than other FTS methods, it still requires heavy computational resource. We recently developed a deep learning approach for the prediction of partial saddle points to enhance the performance of L-FTS. In this presentation, we demonstrate the effect of fluctuation on the diblock copolymer (BCP)-random copolymer (RCP) blend system by comparing L-FTS and SCFT results. Our major focus is on the shift of order-to-disorder transition (ODT) points of Flory-Huggins interaction parameter, χNODT. At low RCP fraction, χNODT predicted by SCFT increases almost linearly as RCP fraction increases. As expected, the χNODT predicted by L-FTS increased faster than that calculated by SCFT, and the degree increases as the fluctuation level goes up.