Bridges constitute key elements of the nation's infrastructure and are subjected to considerable threats from natural hazards including seismic events. A range of potential bridge retrofit measures may be used to limit seismic damage in deficient bridges, and help mitigate associated social and economic losses. However, since resources are often limited for investment in seismic upgrade, particularly in regions of intense but infrequent events, a risk-based approach for evaluating and comparing the cost effectiveness of different mitigation strategies is warranted. This thesis illustrates a method for evaluating the best retrofits for non-seismically designed bridges based on seismic life-cycle costs and cost-benefit analysis. The approach integrates probabilistic seismic hazard models, fragility of as-built and retrofitted bridges for a range of damage states, and associated costs of damage and retrofit. The emphasis on life-time performance and benefits, as opposed to initial retrofit cost alone, not only permits risk-wise investment, but also helps to align upgrade actions with highway agency missions for sustainable infrastructure.