The fatigability of paralyzed muscle limits its ability to deliver physiological loads to paralyzed extremities during repetitive electrical stimulation. The purposes of this study were to determine the reliability of measuring paralyzed muscle fatigue and to develop a model to predict the temporal changes in muscle fatigue that occur after spinal cord injury (SCI). Thirty-four subjects underwent soleus fatigue testing with a modified Burke electrical stimulation fatigue protocol. The between-day reliability of this protocol was high (intraclass correlation, 0.96). We fit the fatigue index (FI) data to a quadratic-linear segmental polynomial model. FI declined rapidly (0.3854 per year) for the first 1.7 years, and more slowly (0.01 per year) thereafter. The rapid decline of FI immediately after SCI implies that a "window of opportunity" exists for the clinician if the goal is to prevent these changes. Understanding the timing of change in muscle endurance properties (and, therefore, load-generating capacity) after SCI may assist clinicians when developing therapeutic interventions to maintain musculoskeletal integrity.