Natural materials such as timber, stone and many others as used in buildings and other structural systems are laden with uncertainties and often prone to risk of failure due to the inherent variability in their engineering properties. To assess the randomness in properties of typical Nigerian grown timbers and propose stochastic models for used in reliability based design of timber structures, a number of timber species were obtained and assessed in the laboratory. Six timber species with respective local and scientific names of Abura (Mitragyna ciliata), Mahogany (Khaya senegalensis), White-Afara (Terminalia superba), Ire (Futumia ebrifu), Madobiya (Pterocarpus erinaceus) and Ayinre (Albizia lebbek), dried to average moisture content of 7.5% were employed for the study. Simple clear specimens of varied dimensions were prepared and tested under uni-axial compressive strength (oriented in parallel and perpendicular to the timber grains) as well as assessment of density and stiffness (modulus of elasticity) in accordance with BS EN 13183-1 (2002), BS EN 408 (1995) and BS EN 408 (2003) using the Cussons universal testing machine of 100 kN capacity at a loading rate of 0.17 mm/s were carried out. Results from the laboratory analysis were further used in proposing the stochastic model using a Matlab toolbox for statistical and simulation analysis where the best stochastic model (that with the least (kolmogorov- Smirnov (K-S) test) is adopted. These, alongside the assessed mode of failure for the timber species reported in the paper would help structural engineers in making informed decisions on the use of these timber materials in design of reliable and safe structures.