The purpose of this study was to investigate the effects of different bandwidth knowledge of results on the acquisition of spatial accuracy and error detection capabilities in a golf putting task. Thirty two university female students (mean age=22.3 years) served as participants and were randomly assigned to one of the followings: BW0% group, BW5% group, BW10% group or BW15% group. Ten-minute and one- day retention tests were administered after experimental treatments. The dependent variables were variable ereor (VE), proportion of correct trials(PCT). correlation of objective and subjective error (COSE) scores. Mixed tow-way ANOVA and Tukey's HSD method showed significant main effects on trail blocks in VE during acquisition. All groups' movement consistency persistently improved with practice. In retention tests, significant main effects on both trail blocks and groups in VE were found. This revealed all groups' movement consistency significantly decreased in the delay retention test, but BW10% group movement consistency was more higher than the other three groups. X �� test was also found that BW10% group had significantly higher PCT than that in BW0% group and BW15% group in the immediate retention test. Moreover, BW10% group also had higher PCT than that in the rest of three groups in the delay retention test. Fisher's z transformation found that both BW10% group and BW5% group had significantly higher COSE than that in the other two groups in the immediate retention test. The results were discussed in terms of blocking hypothesis. Within the limitation of this study, it was concluded that, (1) different bandwidth KR affect the acquisition of spatial accuracy; (2) different bandwidth KR had various effects on the development of learners' error detection capabilities.