Abstract Structural reliability depends on uncertainties in resistance and loads. In many practical cases the resistance dominates and a reduction of uncertainty about resistance is an effective way of increasing safety. It can be accomplished by proof loading. A truncated distribution is considered and reliability indices are calculated for various proof load levels. The structural reliability is sensitive to proof loading for larger coefficients of variation of resistance. A Bayesian approach is applied to develop a posterior distribution for resistance, after proof loading. Reliability indices are calculated for various ratios of the coefficients of variation of load and resistance.