In the prediction of the indoor thermal environment and the heat load, meteorological data obtained at weather stations in each city are usually utilized as meteorological conditions. However, there exists considerable difference in the local climate in each region inside a city. So it is necessary to consider the regional characteristics of meteorological conditions in order to improve prediction accuracy. This study aims to examine the influence of regional characteristics of local climates inside Sendai city on the performance of cross-ventilation in residential houses from the viewpoints of the indoor thermal environment and the cooling load based on the numerical analysis using TRNSYS. Here, the results of numerical analysis of mesoscale climate were utilized as input data of meteorological conditions for TRNSYS.