A residual strength degradation model is presented which deals with both tensile and compressive strength degradation as competing failure modes. Static strength distributions are assumed to follow a two-parameter lognormal distribution, whereas the residual strength degradation equations are based on a three-parameter power law model. A procedure is outlined for determining the parameters in the model. Comparisons are made between theory and experiment for constant amplitude tension-tension and tension-compression fatigue data. Both residual tensile and residual compressive data obtained under these two load histories are considered.