Korea has been entering the ageing society as the population of age over 65 shared 7.2% since 2001. The most urgent issue on the ageing society in Korea is that the speed to entering into the high ageing society is estimated the fastest in the world. It would take 7 years from the ageing society to the high ageing society, which is faster than the period of Japan. The purpose of this study is to find out which factors affect the travel mode choice of the elderly.Multinomial logit model is used to find which factors are able to explain the mode choice of the older people. The pooled model with 711 samples was gathered at Seoul, Gwachon cities and Daegu city for people over 50 years old. The pooled model coefficients were estimated using the explaining variables, such as access time and cost, main travel time and cost, egress time and cost variables with regional dummy variable. In order to find the differences between the age groups and the survey regions the pooled model was segmented as two parts of 50-64 years old and over 65 years old as well as Seoul and Gwachon cities and Daegu city. The model for the over 65 years old showed that older people did not consider time variable significantly when choosing a mode, since the t-value of main travel time variable was low at the 5% level of significance. The test results indicated that older people tends to prefer to choose his/her mode by considering the travel cost rather than the travel time. Most t-values for the estimated parameters was stable at the 5% level of significance. The likelihood ratio index of the model (ρ2) was robust since its value was 0.39 which was known to an excellent value between 0.2 and 0.4. However, this study did not consider jobs or trip purpose when choosing a travel mode by the elderly due to the sample size. These factors which are considered as major factors of mode choice have to be applied in the future study.