The purpose of this paper is to examine whether the five constructs of Senge`s learning organization model is empirically valid in a construction company in Korea and to identify what kinds of learning groups in the company and what different characteristics of these learning groups are existed. The final number of teams used in the analysis is 154. Six constructs were identified by factor analysis and that result seemed to reflect the characteristics of Korean construction industry. Four learning groups were finally drived through multivariate data analysis including cluster analysis, ANOVA and duncan`s multiple range test and each of them showed meaningful differences on the basis of Senge`s five constructs. Some limitations were discussed and future research directions were also suggested.