This study shows that conventional methods in rotor blade fatigue testing can lead to substantial under-testing across major blade areas because material fatigue behavior is not well represented. An improved approach, based on strain-proportional loads with mean load correction, is proposed to define loads that produce sufficient fatigue damage across all blade areas. The results suggest that this can require up to 16 % higher uniaxial fatigue test loads than needed by conventional methods.