Material properties and stresses acting on machine parts are not point values but are statistically distributed. Consequently, it is possible for stress and strength distributions to overlap, exposing weaker materials to stresses greater than their strengths, leading to failure. A Monte Carlo simulation was conducted on AISI 4140 steel subjected to constrained thermal expansion using practical statistical distributions of material properties and environmental variables. The probability of failure of constrained and unstrained AISI 4140 steels were compared and the results showed a lower probability of failure for the latter. The results also showed that at low to moderate stress levels, mechanical load and thermal stress had relatively similar impact on the probability of failure. However, at high stress levels, the effect of mechanical load was predominant.