Considering the uncertainties associated with earthquake, seismic response of structures, damage caused by the response and the repair costs, determining Earthquake Insurance Premium (EIP) is a challenging task that requires a framework for modeling all inherent risks. Most of the models proposed for EIP determination have two basic components. The first component indicates the probability of seismic hazard (generally obtained from the probabilistic seismic hazard analysis), and the second component indicates the probability of occurrence of a given damage to the structure with respect to different seismic hazard levels (generally obtained from the fragility curves). These two components are interconnected through an intermediate parameter, known as the Intensity Measure (IM), which is of great importance in calculations and has hardly been considered so far. In this study, the dependency of EIP on the IM used in the probabilistic seismic hazard analysis and the fragility curves is analyzed. To this end, the EIP was determined for five types of buildings at low, medium and high seismic hazard levels using two IMs (peak ground acceleration and first mode spectral acceleration). The results of this study warn that the EIP can be highly dependent on the IM, so that for all studied structures, the EIP determined based on peak ground acceleration at a high seismic hazard level is nearly three times as many as the one determined by the first mode spectral acceleration. A significant finding is that when the first mode spectral acceleration is used as the IM, the ratio of EIP at different seismic hazard levels closely matches the ratio of IM at the same levels. This result can be valuable in developing insurance codes and regulations. However, this is not the case for the currently used IM, i.e., peak ground acceleration.