The scale of fluctuation (SOF) is used to evaluate the correlation between two locations of geotechnical parameter, which is important for mineral exploration and engineering construction. In this study, the performances of conventional moment methods are firstly estimated, and then the exponential autocorrelation function method that can achieve the best performance in estimating SOF is selected. After that, the Torque Penetration Index (TPI) and the Field Penetration Index (FPI) obtained from the Yinsong Water Diversion Project are used to detect the scale effect of SOF and to quantify the influences of rock types and surrounding rock mass grades on SOF. As the spatial scale increases, the SOF values increase, and the uncertainty in the SOF of the surrounding rock mass properties increases. The SOF calculated using TPI is slightly higher than using FPI. The mean values of SOF for granite, tuff, diorite, and limestone are quite close at the 50 m–200 m scales, with coefficients of variation (COVs) of SOF range from about 0.2 to 0.4. There are differences in SOF between grade III and grade IV surrounding rock masses at the 50 m–500 m scales. The mean values of SOF in grade III surrounding rock mass are about 4%–15% higher than that in grade IV surrounding rock mass, and this difference in general increases as scale increases. The rock mass with better conditions has a stronger spatial correlation. The COVs of SOF for the two rock grades reached about 0.3–0.5, owning to the inherent spatial variability.