This article examines the extent to which post-weld treatment by needle peening can improve the fatigue performance of tubular truss bridges. To do this, the various potential crack sites on several variants of a typical bridge are analyzed using a probabilistic, fracture mechanics-based model. Systems reliability theory is then used to determine the reliability of the entire untreated or treated bridge. The results of this work show that: considering phase effects may result in large reductions in the design stress ranges for these structures, a significant increase in the treatment benefit can be achieved if the treatment is applied after the dead load stresses are introduced, and weld root cracking does not appear to be the critical failure mode for these structures, so long as a strategy of partial treatment is employed.