This thesis presents an analytical study of the behavior of prestressed composite beams. The particular configuration of prestressed composite beam considered is a simply-supported steel beam, prestressed by a high-strength steel tendon with a constant eccentricity, and attached to a concrete slab by shear connectors designed to insure complete interaction. A numerical method for analyzing the statically indeterminate beam was developed and used as a basis for a computer program which found the strains and deflections at a discrete number of equally spaced points along the length of the beam for increasing values of load up to the failure load. These strains and deflections satisfied the stress-strain relations for the materials, the static equilibrium of the beam, and the compatibility of deformation of the tendon. Hognestad's stress-strain relation was assumed for the concrete slab, an elasto-plastic stress-strain relation was assumed for the steel beam, and an actual stress-strain curve was used for the tendon. The numerical procedure combines the method of tangents for solving simultaneous nonlinear equations with a method of successive approximations. A simplification of this program was devised to analyze conventional composite beams. Detailed studies were made of some prestressed composite beams suitable for highway bridges and some suitable for buildings. The effect on behavior of variation of prestress force and tendon size were investigated and compared with the effect of variation of cover plate size for conventional composite beams. These studies show that prestressing a composite beam is an effective means of increasing the load capacity of the beam at all of the following stages of behavior: load causing allowable steel stress, load causing yielding of steel beam, and ultimate load. Prestressing does not significantly increase the load causing the allowable concrete stress nor does it significantly reduce live load deflection. The behavior of a prestressed composite beam is shown to be not very sensitive to variation of slab thickness. The validity of the method of analysis was verified by the comparison of analytical and experimental results for several prestressed and conventional composite beams which were tested in the laboratory by other investigators.