Abstract: The earthquake return period is one of the important concepts in earthquake engineering, which has been widely used in seismic analysis and evaluation of structures. Nowadays, the earthquake return period concept is usually based on one intensity measure, which cannot include joint and conditional distribution information of ground motion intensity measures. The concepts of bivariate and conditional earthquake return periods are presented, and the basic theories of bivariate and conditional earthquake return periods are given. For the case site, vector and conditional probabilistic seismic hazard analysis ar“e conducted, and bivariate and conditional earthquake return periods are obtained, which are used for generating vector and conditional site-specific spectra. Theory and implementation prospects of bivariate and conditional earthquake return periods are given. The results show that bivariate and conditional earthquake return periods include correlation information of intensity measures on basis of single variable earthquake return period; bivariate earthquake return period value is larger than or equal to a ny single variable earthquake return period of two intensity measures, and conditional earthquake return period is the ratio of bivariate earthquake return period over single variable earthquake return period; bivariate return period surfaces and contour lines are different for different combinations of intensity measures, which are usually related to the correlation coefficient of two intensity measures and the hazard level of intensity measures; the smaller the conditional intensity measure, the greater the conditional earthquake return period of the same predicted intensity measure.