The performance of regenerated cellulose (RC) films and capsules was investigated for their applications in oral controlled drug delivery. Regenerated cellulose films were prepared by non-solvent-mediated, phase inversion of native and depolymerized cotton linter solutions (methylolcellulose; cellulose dissolved in dimethyl sulfoxide/ paraformaldehyde solvent system) in water as well as by phase inversion of native cotton linter solutions in organic non-solvents followed by thermal annealing. These films were monolithic in dry state and formed porous structures when hydrated. Irrespective of the degree of polymerization of the starting cellulose source or the use of organic non-solvents, the cellulose chain length was not significantly altered and cellulose was in an amorphous state. Flux analysis in diffusion cells, using ethanol-water mixtures as the solvent medium, indicated that the films take up solvent to form porous routes for transport of solute. The amount of solvent uptake required to form these routes was greater for films prepared from depolymerized cotton linter. Ionic and hydrophobic solutes traverse the films using the porous pathways following hydration of the film. Blended RC films were prepared by combining native and depolymerized cotton linter solutions in varying ratios and phase-inverting in water, followed by thermal annealing. Porosity, pore size and water uptake of the hydrated films decreased, while the length of the transport pathway (tortuosity) increased, as the fraction of depolymerized cellulose increased in the blended films. Differences in methylene blue dye adsorption on phase-inverted vs. phase-inverted and thermally annealed RC films indicated that the type of non-solvent utilized for phase-inversion does not affect the internal RC film structure during the phase-inversion process. However, as the boiling point of the non-solvent increased, the amount of irreversible polymer consolidation and formation non-swelling domains (hornification) increased during the thermal annealing process. This, in turn, led to reduced porosity and solute flux through these RC films. Two-piece cellulose capsules were fabricated by phase-inversion of methylolcellulose solutions in water using a dip-coating approach. Zero-order release rates for a number of drugs increased as their water solubility increased. The release of water soluble drugs occurred by osmotically-driven convection and diffusion through the pores in the capsule wall, while the release of moderate to poorly soluble drugs predominantly occurred by diffusion. Moreover, as the drug solubility increased, the apparent permeability of the drugs through the capsule wall decreased, which indicated that the inward osmotic flux of water reduced the diffusivity of the drug through the pores. The hydraulic permeability of the cellulose capsules was determined to be higher than for conventional ethylcellulose and cellulose acetate coated osmotic drug delivery systems, indicating that the cellulose-based capsules may be better suited for osmotic drug delivery.