The purpose of this study was to examine the reliability and validity of the observed behaviors of subjects who suffered from low back pain with spinal diseases, Thirty two low back pain patients admitted on the neurosurgical unit in an army hospital were compared with 30 normal controls belonged to an army unit, by means of matching the age, hight and weight. Observed pain behaviors were developed by the researcher on the bases of literature and patient observation. This tool consists of 18 behaviors seperated into two major groups : mutually exclusive and concomittent behaviors. The mutually exclusive behaviors included coding cathegories for 6. body motions assumed by the subjects during the observation session. These 6 standardized motions consisted of sitting from standing first, and serially tying down, reclining, sitting again, and then standing, 6 steps walking. Concomittent behaviors consisted of 12 observable patterns that can be observed systematically from the face, grimacing, bracing, rubbing, walking with arms fixed, support with hands on sitting or standing, guarded movement, limping, unbalaced weightbearing, stopped movement from tying position to sitting, sighing and graoning. Subjects were videotrecorded as they performed a 6-standardized sequence of motions, simultanously researcher measured the time spent performing each motion and step length. Patients were asked torate their subjective pain score on the 10 mm graphic rating scale ranging from 'no pain' to 'sever pain'. For scoring of the pain behaviors, two trained nursing officiers independently and simutanously viewwd each videorecording and checked subject 'pain behaviors at the observational item checklist. The result of the study are summarized as follows : 1. Reability of the observational tool was a=.845. 2. Spearman's rho and percentage agreement were p=.97 and 81.7 persent respectively, that indicate adequate interrater reability of this tool. 3. The sensitivity rate of the tool was .875 while specificity rate .866 for differentiating patient from the normal. 4. When difference in the objective pain indices between patient group and control were compared, there was significant difference of all indices, such as pain behavior(t=7.71, p=.0001). spent time performing motion(t=14.2, p=.0001), step length (t=-10.72, p=.0001). 5. There were differences in the objective indices the subjective pain subgroups (low, medium, high). Differences in the mean score of objective pain behavior (F=6.376. p=.005) and spent time for moyion(F=4.631, p=.018). But there were no significant differences in the step length among the subgroups(F=.667, p=.521). 6. Highly correlated pain behavior items wiyh subjective pain score were 'stopped movement from lying position to sitting', 'limping', 'support with hands on sitting or standing', 'bracing', 'guarding' and 'walking with arms fixed'. In summary, although some of rho behavior items such as sighing and groaning in this study could not be observed because of videotaped datd, the reliability and validity of the over all observation method were satifactory. Thus, the results of the present study demonstrate rye potetional utility of the tool in assessing objective pain complementing self-reported pain in low back pain patients.