Based on the standards for design of structures, any structure should be designated for seismic loads and any combinations containing seismic loads. For spectral analysis of a structure, the site effect is taken into account by considering its effects on the design spectra. Because of different lateral and vertical stiffness of the soil layers underneath the structure, the design spectra are different from soil to soil. The stiffer the soil the higher the velocity results in. The shear and compressive wave velocities of the soil as continuum are criteria for categorizing the soil in stiffness point of view. Since, the structures are more affected by lateral forces than vertical ones, the shear wave velocity is more important than the compressive wave velocity. Moreover, the soil nearer to the structure affects the structure more than the soil far from it. Thus, in the standards for design of structures, the mean shear wave velocity of the upper 30 m of the soil layer is used for categorizing the soil underneath the structure in stiffness point of view. In the Iranian code of practice for seismic resistant design of buildings, the standard number is 2800, the sites have been categorized into four different types, which are rock (with the mean shear wave velocity of the upper 30 m denoted as larger than 750 m/sec), medium alluvium (where m/sec), soft alluvium (where m/sec), and very soft alluvium (with m/sec). To analyze the structures using design response spectra, specified horizontal and vertical spectra are needed for each category. Because of the increase in the number of strong motion accelerograms in recent years, in this research the horizontal and vertical design spectra for the first category (the rock site) are prepared based on Iranian data. To do so, all the existing horizontal and vertical acceleration time histories in different stations fixed on rock sites are gathered. The data are filtered and baseline corrected by Seismosignal software to remove the noise frequency components and to modify the magnitude of both the displacement and velocity. In addition, all the data are normalized for the Peak Ground Acceleration (PGA). According to the data, the quality of 60 vertical time histories and 71 horizontal time histories were acceptable. With this data, both the vertical and the horizontal response spectra are prepared for each time history and for four different damping ratios, which are %2, %5, %10 and %20. Averaging the response spectra, the unsmoothed design spectra are obtained. The smoothed design spectra are plotted in tripartite coordinate system and spectral acceleration-time system, as well. These procedures are done for the average plus one standard deviation of vertical and horizontal response spectra. Finally, the smoothed design spectra from the data of this research are compared with that of the Iranian code of 2800 regulation and also the Mohraz design spectra. It is shown that the results are in good agreement with the Mohraz design spectra except that in long periods, the spectral acceleration obtained in this study is smaller. Comparing the result of this research with that of 2800 regulation, it is seen that in short periods, the spectral acceleration in this study is higher than that in the 2800 regulation, while for long periods, the spectral accelerations in this study is much less than that given in 2800 regulation. It means that in the category of short period structures more strengthen structures may be needed.