99% of the giant panda′s (Ailuropoda melanoleuca) diet consists of the subalpine bamboo species. Thus, the life span of the giant panda is directly related to the life span of the bamboo it feeds upon. The regeneration of bamboo is linked directly to the giant panda′s survival and conservation. Our goal was to find out whether giant panda grazing or human harvesting would best promote clonal regeneration of bamboo. From 2003 to 2007, using the plot sample and positioning observation methods, clonal regeneration of umbrella bamboo (Fargesia robusta) was studied in the Wolong Natural Reserve of China. Three types of plots were located in the reserve: grazed (bamboo in the wild training enclosure from July 2003 to Sept. 2004), harvested (bamboo cut in the spring of 2004), and control. Results showed that under similar demographic and environmental conditions (p>0.05), the grazing and death rates of bamboo for the grazed plot were 67.07% and 29.07%; vs. the harvested plot of 65.67% and 46.68%. Compared to giant panda grazing, harvesting strongly restricted the viability of the umbrella bamboo population. The result of analysis on the bamboo shoots indicated that grazing and harvesting promoted the shooting rate of umbrella bamboo (p<0.05). The quality of bamboo shoots (length and diameter) in the harvested plots was no better than that of the grazed and control plots (p<0.05), and giant pandas were inclined to feed on the bamboo shoots in the harvested plots (p<0.05). The fecundity of the bamboo shoots varied over the years in the grazed and harvested plots (p=0.006-0.035). After a long period of renewal, the densities of the bamboo shoots converged in all plots. Apart from 2007, annual culm recruitment of the umbrella bamboo clone population was significantly different among the three plot types from 2003-2006 (p<0.05). Annual culm mortality was significantly different among the three plot types during 2003 and 2004 (p<0.05), but no significance from 2005 to 2007. In the grazed and harvested plots, the mortality rate was significantly lower than the recruitment rate in 2003, 2004, 2006, and 2007 (p<0.05). However, the mortality rate of the grazed plot in 2004 and 2005 along with the harvested and control plots in 2005 showed an increase in the recruitment rate. There was no significant difference between the recruitment and mortality rates (p>0.05). Overall, both grazing and harvesting measures had the potential to improve the umbrella bamboo fecundity, and results suggested that clonal regeneration of umbrella bamboo was actively affected by giant panda grazing and proper harvesting.