Energy-based design and assessment approaches usually require use of ground motion datasets. Previous researchers attempted to use solely actual records, however, simulated records can also be used as an alternative in the absence of a homogenous ground motion database. The main objective of this study is to examine the differences in energy-based responses of structures subjected to real and simulated datasets. For this purpose, the obtained input energy spectra based on 'simulated' and 'real' records are compared. The simulated dataset is generated by the stochastic finite-fault method, considering the regional seismicity parameters that correspond to Duzce city located on the western segments of North Anatolian Fault zone in Turkey while the real dataset is formed by selection of regionally compatible records from NGA-West2 ground motion database. Next, time history analyses are performed for both elastic and inelastic SDOF structural systems to investigate the differences in terms of input energy, resulting from alternative ground motion datasets. The analysis results present variability in terms of energy-based structural demand parameters when alternative ground motion datasets, real or simulated records, are employed.