Dynamic models for RC frame structures with 4-floor, 8-floor, 12-floor and 16-floor covering common period ranges were established and verified to systematically investigate effects of hanging wall/footwall fault parameters on dynamic responses of RC frame structures. Three NGA ground motion attenuation relation models including ASK, CB and CY models were adopted to fit hanging wall/footwall earthquake records for purposes of comparison and optimization. The optimal fitting model was determined through comparison between fitted results and real 622 hanging wall/footwall ground motion records. Finally, 160 hanging wall/footwall earthquake samples with different fault parameters were fitted based on the optimal NGA model. The relations between hanging wall/footwall ground motion PGA and parameters, such as, earthquake magnitude, soil body shear wave velocity, fault dip angle and upper bound embedded depth were investigated. The effects of hanging wall/footwall fault parameters on dynamic responses of four RC frame structures were studied. The results showed that ASK model can simulate hanging wall/footwall earthquakes well and its simulated results agree best with real earthquake records; PGAs of hanging wall/footwall earthquakes increase with decrease in site distance and increase in earthquake magnitude; fault dip angle has a larger influence on PGAs within hanging wall site distance of 20 km; soil body shear wave velocity and upper bound embedded depth have little effects on PGAs; dynamic responses of four RC frame structures increase with decrease in site distance and soil body shear wave velocity, they increase with increase in earthquake magnitude and structure height; fault dip angle and upper bound embedded depth have almost no effect on structural dynamic responses under a certain footwall site distance; hanging wall earthquakes have a larger effect on dynamic responses of lower RC frame structures, while footwall earthquakes have a bigger effect on dynamic responses of higher ones.