The accuracy of two methods of non-destructively detecting wood decay in living trees was compared, a stress-wave velocity measurement and a sounding test using a hammer. Measurements were conducted on trees in a 53-year-old Larix kaempferi stand. In both measurements, the accuracy of detecting wood decay in living trees increased as the percentage of decayed diameter in the total diameter of the trunk increased. The accuracy of detecting decayed trees by stress-wave measurement was higher than that of the sounding test when the decayed diameter exceeded 30 percent of the trunk diameter. However, both measurements yielded low accuracy when less than 30 percent of the trunk diameter was decayed. The stress-wave velocity varied greatly even in sound trees, and this decreased the accuracy of the detection by stress-wave measurement, especially in trees, which were less decayed.