In the field of new economic geography, recent years have produced new empirical attempts to investigate the wage structure resulting from regional potentials. Pioneering studies are reported by Hanson (2005) , Redding and Venables (2004) and Brakman et al. (2006) . Although the model structure is somewhat complicated, the idea is simple and familiar because of its relation to traditional gravity analysis. Regional potentials are determined by market access and supplier access. In regions with a large market or numerous suppliers, firms have advantages in competition over firms in other regions. Employees in such regions receive higher wages. This paper constructs a new economic geography model of Japan, and takes a simulation approach to investigate the transition of regional potentials and labor distributions with trends of a decrease in transportation costs. We first calibrate the main parameters (including the elasticity of substitution, transportation cost levels, etc.) , and derive the regional potentials in Japan. Then we predict (and replicate) the transition of regional potentials and labor distribution by changing the transportation costs. Consequently, we show that (i) the simulation result approaches the actual labor distribution, (ii) decreased transportation costs stimulate the dispersion of economic activities, and (iii) the simulation replicates the following the historical movement of the economic center of Japan. Some time ago, economic activities were concentrated in the Kansai area (including Kyoto and Osaka) when transportation costs were extremely high, but these activities were displaced to the Kanto area (including Tokyo) as transportation costs decreased. JEL Classification: R11, R12, F12