The objective of this thesis is to develop knowledge and methods to prevent fatigue failure of bridges. The Fatigue Limit State is investigated from both the demand (S) and resistance (R) side. The S-part focuses on reduction of uncertainties through monitoring of fatigue action effects: •Analysis of 28-month-long high-frequency monitoring of a slab’s portion of a prestressed concrete road bridge demonstrated that the stress ranges due to traffic loading and temperature action can be of similar magnitudes. Furthermore, they should be treated together as their combination can be fatigue-relevant. •The monitoring duration influences the reliability of results, which are further extrapolated to obtain the Cumulative Fatigue Damage for the total service duration of a structure. To quantify the associated uncertainty, the monitoring duration-dependent Cumulative Damage correction Factor (CDF) was calibrated. The minimum recommended monitoring duration is 100 days, and for this observation period the correction factor γCDF=4 for massive structures and γCDF=20 for temperature-sensitive elements should be applied. After one year of monitoring γCDF can be reduced to 1.3 and 2.5 respectively. The correction factor is significantly smaller than the one obtained with method suggested by Eurocode, leading to γCDF=20. The R-part of this thesis concentrates on a structural response and fatigue resistance of reinforced UHPFRC (R-UHPFRC) and was based on experimental testing of full-scale R-UHPFRC beams. The results can be applied to both new structures and elements strengthened with R-UHPFRC layer: •It is demonstrated that after loading-unloading cycles, due to modified mechanical properties of a part of UHPFRC element which entered into strain-hardening domain, the distribution of stress in the cross-section is re-arranged, influencing the global structural response. Two important fatigue-relevant conclusions are drawn: I)the stress range in the rebar is much lower than calculated using initial material properties; and II)the portion of UHPFRC is subjected to tensile-compressive rather than to tensile-tensile fatigue stress. •The fatigue phenomenon of R-UHPFRC member is observed in detail using Distributed Fibre Optics sensing. The strain variation during the fatigue process in both UHPFRC and rebar remained stable for most of the experiment. The rapid increase of strain range occurred at around 90% of the test duration. During the last 1% of the fatigue test, the increase of strain in the reinforcement bar took place while the increase of the beam deflection range ensued during last 1‰ of test duration. Importantly, the failure of reinforcement bar is identical with failure of the member. •Fourteen beams were exposed to constant amplitude fatigue in four-point bending. It was the largest experimental campaign on R-UHPFRC ever executed. Two-level fatigue verification was proposed: I)global verification using normalized minimum and maximum load levels and the modified Goodman diagram, contrary to previous methods based on the maximum load and load range; and II)local verification of stress-range in reinforcement bar using the standard S-N curves. The research work presented in this thesis brings new knowledge in both fatigue demand (S) and resistance (R) of both R-UHPFRC structures and reinforced concrete structures strengthened using UHPFRC. More economic solutions for structures can be obtained, leading to both financial and environmental savings.