For the progress of durability and service life of wooden constructions through appropriate maintenance, it is important to detect biodeterioration and to evaluate residual strength of deteriorated wooden construction members. To establish reliable and practical nondestructive methods to estimate residual bending strength (RBS) of biodeteriorated wood, we examined the adaptability of the method of stress-wave propagation: 1) the effect of patterns and degree of deterioration on stress-wave propagation; and 2) relations between stress-wave propagation and RES on wood specimens with many patterns of artificial defects as models of biodeterioration. We used 41 small-size wood pieces and 65 construction-size timbers. We processed 12 pat-terns of artificial defects by one or two "slits" or one "ditch." As estimation parameters of RES, we adopted stress-wave propagation time (t-sw), modulus of elasticity by stress-wave propagation time (E-sw), stress-wave propagation energy (Eg-sw), and attenuation of stress wave (At-sw). 1. At distances between two loading points across defects, the possibility to detect defects by t-sw was observed based on an upward tendency of the ratio of t-sw after process to that before process with an increasing defects depth percentage. 2. At distances between two loading points across defects, a high possibility to estimate RES by t-sw was observed. 3. On construction-size timbers, a possibility to estimate RES by E-sw was observed at three distances across defects. Correlations between E-sw and RES were smaller than those between t-sw and RES. 4. The possibility to detect defects by Eg-sw and At-sw, and the capacity for advantage of Eg-sw and At-sw as evaluation parameters for RES were shown at distances between cross sections at ends of specimens and between two loading points across defects.