Natural fiber-plastic composites incorporate thermoplastic resins with fibrous plant-based materials, sometimes referred to as biomass. Pine wood mill waste has been the traditional source of natural fibrous feedstock. In anticipation of a waste wood shortage other fibrous biomass materials are being investigated as potential supplements or replacements. Perennial grasses, agricultural wastes, and woody biomass are among the potential source materials. As these feedstocks share the basic chemical building blocks; cellulose, hemicellulose, and lignin, they are collectively called lignocellulosics.;Initial investigation of a number of lignocellulosic materials, applied to fiber-plastic composite processing and material testing, resulted in varied results, particularly response to processing conditions. Less thermally stable lignocellulosic filler materials were physically changed in observable ways: darkened color and odor.;The effect of biomass materials' chemical composition on thermal stability was investigated an experiment involving determination of the chemical composition of seven lignocellulosics: corn hull, corn stover, fescue, pine, soy hull, soy stover, and switchgrass. These materials were also evaluated for thermal stability by thermogravimetric analysis. The results of these determinations indicated that both chemical composition and pretreatment of lignocellulosic materials can have an effect on their thermal stability.;A second study was performed to investigate what effect different pretreatment systems have on hybrid poplar, pine, and switchgrass. These materials were treated with hot water, ethanol, and a 2:1 benzene/ethanol mixture for extraction times of: 1, 3, 6, 12, and 24 hours. This factorial experiment demonstrated that both extraction time and medium have an effect on the weight percent of extractives removed from all three material types.;The extracted materials generated in the above study were then subjected to an evaluation of thermal stability by thermogravimetric analysis in a subsequent experiment. Overlay plots, combining individual weight loss curves, demonstrate that the experimental factors, solvent system and extraction time, produce effects on the thermal stability of the treated biomass samples. These data also indicated that the individual lignocellulosic materials had unique responses to the type of solvent used for pretreatment. Increasing extraction time had either no correlation with or a positive effect on thermal stability of the biomass samples.