The numerical approach to probabilistic key block analysis is used to evaluate the influence of joint orientation on key block size distribution. A series of tests are conducted by variation of three joint sets. Since joint orientation is recognized by the dispersion of pole vectors (k-factor) so that they are varied. The other joint parameters are kept constant during the test. The relative angle between pole vectors (A) is tested at 90, 60, 30 and 0 degrees from each other. The concentration factor of pole vectors (k-factor) is varied for each different angle. The distribution of key block size is observed from each test. The results show that a unimodal-shaped distribution occurs when k-factor is reduced from 10,000 to 200. When joint plane orientations are random or k-factor is less than 200 the resulting key block size distribution fits a reverse J-shaped Weibull distribution. The fitted shape parameter of Weibull distribution reduces as k-factor is reduced. The angle between pole vectors has less influence on the key block size distribution than does the pole dispersion.