The objective of this study was to determine how mechanical stimulation affects the biomechanics and histology of stem cell–collagen sponge constructs used to repair central rabbit patellar tendon defects. Autogenous tissue-engineered constructs were created for both in vitro and in vivo analyses by seeding mesenchymal stem cells from 10 adult rabbits at 0.14×10⁶ cells/construct in type I collagen sponges. Half of these constructs were mechanically stimulated once every 5 min for 8 h/day to a peak strain of 4% for 2 weeks. The other half remained in an incubator without mechanical stimulation for 2 weeks. Samples allocated for in vitro testing revealed that mechanically stimulated constructs had 2.5 times the linear stiffness of nonstimulated constructs. The remaining paired constructs for in vivo studies were implanted in bilateral full-thickness, full-length defects in the central third of rabbit patellar tendons. Twelve weeks after surgery, repair tissues were assigned for biomechanical (7 pairs) and histologic (3 pairs) analyses. Maximum force, linear stiffness, maximum stress, and linear modulus for the stimulated (vs. nonstimulated) repairs averaged 70% (vs. 55%), 85% (vs. 55%), 70% (vs. 50%), and 50% (vs. 40%) of corresponding values for the normal central third of the patellar tendons. The average force-elongation curve for the mechanically stimulated repairs also matched the corresponding curve for the normal patellar tendons, up to 150% of the peak in vivo force values recorded in a previous study. Construct and repair linear stiffness and linear modulus were also positively correlated (r = 0.6 and 0.7, respectively). Histologically both repairs showed excellent cellular alignment and mild staining for decorin and collagen type V, and moderate staining for fibronectin and collagen type III. This study shows that mechanical stimulation of stem cell–collagen sponge constructs can significantly improve tendon repair biomechanics up to and well beyond the functional limits of in vivo loading.