The authors have been proposed and developed snow-melting system using geothermal and solar energy. In summer, solar heat is stored into underground from road surface to underground piles. In winter, the underground heat is utilized to melt snow on the road surface. This system was applied to parking lots and bridges of relatively small scale (less than 1000 m2 ). Numerical simulation program was also developed to predict temperature field of the system and to evaluate system performance. This program was verified by experimental data only for relatively small scale test area. In addition, appropriate design conditions, such as pile diameter, length and number, can not be easily estimated when road surface area and ability (average heat flux) of snow-melting are given. This paper aims to demonstrate the system for relatively large scale (larger than 1000 m2 ), and to obtain optimal design conditions of the system at given road surface area and ability. The snow-melting system using geothermal and solar energy was applied to a parking lot and a bridge of large scale. Both sites were under practical use which means cars were sometimes parked and run over the bridge. Obtained experimental data of temperature field of the system and snow melting situation show that numerical simulation program predicted system performance and temperature field adequately even though the program contains several simplifications. To discuss the optimal design conditions, numerical simulation was conducted by changing the following parameters: diameter, length, number and pitch of piles, pitch and diameter of heat dissipation pipe, flow rate of circulating water, road surface area. All these parameters are considered to affect system performance. The simulation results revealed that pile surface area determined by diameter, length and number of piles is the dominant parameter for deciding snow-melting ability. Namely, when road surface and snow-melting ability are given, necessary pile surface area can be obtained from the simulation results, and system design of piles becomes possible with considering cost for embedding piles.