This fundamental study analyzes the relationship between rock thermal properties and psychrometric properties in underground space and has a ultimate goal to develope technologies for predicting major environmental variables. The study is divided into 2 subjects (1) developement of a basic model for predicting temperature and humidity, (2) analysis of the validity of the model through application to a local underground storage space for military supplies. The basic model is built for the network of tunnel-shaped underground spaces. The model takes into account rock thermal properties and changes in moisture content in the air due to condensation/evaporation on the rock surface. Using lumped-parameter analytical method, heat flux from or to the surrounding rock is calculated and then the psychrometric properties(air quantity, pressure, temperature, humidity) are estimated through network simulation. The model can be utilized regardless of the tunnel type. The study site is a local storage space built in rock, mainly granite gneiss and quartz-porphyry. It is a U-shaped tunnel, 593.5m long and 6x6.5m wide. Relative humidity inside has to be strictly controlled under 55% to avoid erosion of a certain types of supplies stored in 6 chambers with the capacity of 300~1.000 ton. The thermal conductivity varies between 2.734 and 2.779W/m$^{circ}C$ and the thermal diffusivity is in the range of 1.119 and $1.152{ imes}10^{-6};m^2/s$ the specific heat between 910 and $920;J/kg^{circ}C$. Relative errors of the predicted values of dry/wet temperature and relative humidity are 0.8~3.0%, 0~7.5% and 0~7.0%, respectively. Apparent errors associated with the rock surface temperature seems to be partly due to the intrinsic limitations in the infrared thermometer used in this study.