Disease or mortality rate map, geographically is desired for public health officials and epidemiologists. In the classical approach, rates are estimated under the assumption of independence between geographical areas. However, this assumption is not substantiated in health studies, particularly in the case of health variables which may be strongly correlated due to spatial correlation in adjacent areas. In this paper we account for such a dependence. This paper presents Poisson kriging which is now available and applicable to the spatial interpolation of area disease rates. The aim of this research is to improve the precision of parameters estimation for gastrointestinal cancer incidence rate in Iran using Poisson kriging. The methodology is illustrated using gastrointestinal cancer data recorded in ministry of health and medical education (in the non-infectious diseases management center) of Iran on years 1382-1386 to the 336 counties.