The processing of solid wood generates a large amount of waste, which alternatively to disposal and burning in the open air, these can be used in the manufacture of wood products. Among the various stages of production of particleboards, the homogenization process requires consumption of energy, time and labor, and its eventual suppression would certainly result in a reduction in the cost of manufacturing these materials. As there are several properties resulting from the characterization of panels, obtained in equipment generally available in large research centers and branch companies, the relation of properties through mathematical models makes it possible to reduce the volume of tests, as recommended by the Brazilian standard [1]. This research aimed, with the use of a mix of wood shavings (fine particle formed in processes such as planing and thinning wood) in the integral form (without dimensional classification) of Pinus elliottii and Pinus taeda woods (12% moisture content) and of the urea-formaldehyde adhesive, to evaluate the feasibility of producing medium and high density particleboards, the influence of density (medium and high) of composites on the physical and mechanical properties as well as evaluating the possibility of estimating properties as a function of others by linear regression models, also considering colorimetric parameters. Six medium and six high density particleboards were produced considering the use of 15% adhesive content, with the proper use of only 11% adhesive. The physical and mechanical properties were obtained according to the assumptions and calculation methods of the Brazilian standard [2], with the requirements being evaluated based on this and some international standards. In general, the density of the panels was significant in practically 50% of the determined properties. The particleboards can be classified as P2 by the Brazilian standard [2], it should be noted that in some properties the values exceeded the P7 class. The results of thermal conductivity show the potential for application of the panels in buildings. The surface roughness was considered intermediate (class N7 of Brazilian standard [3]). From the regression models, only four of the twenty generated had a coefficient of determination close to 70%, however, because they are all considered significant, a greater number of samples and experimental conditions should be considered for more robust conclusions, should be the objective of future research.