In conventional methods used to model spatial survival data and the factors affecting it, it is often assumed that the coefficients of explanatory variables in different regions of the study area have the same effect on survival time, and usually the spatial correlation of data is considered in the model through a random effect. However, in many practical problems, the factors affecting survival time and relative risk do not have the same effects in different regions. In this paper, the effects of factors affecting relative risk that are not the same in different regions are examined. For this purpose, spatial regression and spatial variable coefficient models are introduced and the Bayesian estimation of their parameters is presented. Then, three regression models, classical regression, spatial regression, and spatial variable coefficient regression, are used to model the factors affecting the excess relative risk of esophageal cancer patients, and the risk caused by different factors is examined and evaluated.