Traditionally, programming of shape memory polymer (SMP) material requires initial heating above the glass transition temperature (Tg ), subsequent cooling below Tg and removal of the applied load. Therefore, the shape fixity process is inconvenient for some applications. Most recently, a new and effective approach, which programs glass transition activated SMPs directly at temperatures well below Tg ,was introduced by Li and Xu [2011, “Thermomechanical Behavior of Shape Memory Polymer Programmed at Glassy Temperature: Testing and Constitutive Modeling,” J. Mech. Phys. Solids, 59 (6), pp. 1231–1250. The 1D compression programming below Tg and free shape recovery were extensively investigated both experimentally and analytically. The current work extends this study into a shape memory polymer based self-healing syntactic foam, which was found to be capable of self-sealing structural scale damage repeatedly, efficiently, and almost autonomously [Li and John, 2008, “A Self-Healing Smart Syntactic Foam Under Multiple Impacts,” Compos. Sci. Technol., 68 (15–16), pp. 3337–3343.]. A structural-relaxation constitutive model featuring damage-allowable thermoviscoplasticity was then developed to predict the nonlinear shape memory behavior of the SMP based syntactic foam programmed at glassy temperatures. After validated by both 1D (compression) and 2D (compression in longitudinal direction and tension in transverse direction) tests, the constitutive model was used to evaluate the effects of several design parameters on the thermomechanical behavior of the SMP based syntactic foam. It is concluded that the model is a useful tool for designing and training this novel self-healing composite.