Abstract: The purpose of this study was to identify the autocorrelation spatial variations of heat Islands of Tehran. For this purpose, network database of minimum and maximum temperature of Tehran was created. Then, the database was used to consider a 33 years statistical period in daily interval of 1/1/1983 to 31/12/2013. Based on the present study and a cell with dimension of 18×18 Kilometres was developed on under study region. The modern methods of spatial static such as global method of Moran spatial autocorrelation, and local Anselin Index of Moran and Hot spots were used to gain with year changing by programming software of GIS. The results of the study showed that spatial variations of heat Islands of Tehran have up word cluster pattern. Among them , according to the local Index of Moran and Hot spots heat Islands have positive spatial autocorrelation pattern (hot heat Islands) in East and Northeast of Tehran and west and South west parts of Tehran have negative spatial autocorrelation pattern (cold heat Islands). In most cases, a large part of province in almost half of total area of province didn’t have any significant pattern or spatial autocorrelation during under study period. Moreover, it was founded that heat Islands are generated and controlled by affecting from two systems: 1- local factors controlling the place (geographic arrangement of Islands), 2- external factors controlling the time (the region of heat Islands).