The conventional KINS2 model combined TRILUX model and response matrix method with the Becker type thermal leakage correction model has been found to be inefficient in the computing time and accuracy when applied to the realistic PWR``s core analysis. It is attributed to the fact that the analytical algorithms for the transport parameter evaluation requires the effective multiplication factor isolation and the parameter transformation and thus causes to increase the computing time as the system size becomes larger. The other fact is that the thermal transients by the Becker type model at the interface of a three dimensional node tend to be overestimated due to the simple summation of leakage effects evaluated in one dimensional calculation using the asymptotic thermal-to-fast flux ratios in a node center. Two methods to overcome these flaws were investigated and developed in this thesis. In order to reduce the computing time, the compact response matrix formalism was developed by introducing the method of variable separation and an assumption into the analytical algorithms. To improve the calculational accuracy, the effect of transverse neutron leakage across the surface perpendicular to the calculational direction was accounted for more accurate thermal-to-fast flux ratios at the node center in the one dimensional calculation. These two methods developed in this thesis has been incorporated into KINS2 and applied to various test problems and to Korea Nuclear Unit 1 and 7 (KNU-1 KNU-7). All the results except for the case of thermal transients being negligible show that these methods rapidly predicts core design parameters in an more efficient manner, especially in case of checker-board geometry such as power reactors. In KNU case, the computing time is saved by a factor of two and the relative power distribution is more accurately predicted by about two in terms of standard deviation of \% difference compared to the conventional KINS2 method.