This study develops autocorrelation spatial model as a software for mapping high risk dengue fever transmission area. the data used in this study are the cases of dengue fever reported to the health department in salatiga from 1998 to 2008. it also uses the number data of mosquito larva free of salatiga. the data analysis method used in this study s global autocorrelation represented by moran's index and local indicator spatial association (LISA) represented by cluster map and significance map. the result of the study shows that the moran index is 0.0251 and -0.0049 which describes the association between vaeiables which is negative towards population density variable and mosquito larva free number (abj) towards dengue fever case. hotspot (positve spatial association area) i salatiga and gendongan have an impact towards the hotspot formation in mangunsari, sidorejo lor and kutowinangun in 2006, kelurahan sidorejo lor and tegalrejo in 2007; mangunsari and gendongan 208. the result of the sudy using LISA has not described that dengue fever sace is always significant towards the increase of abj variable value. based on the map of high risk dengue fever transmission modelled with spatial autocorrelation, then preventive and anticipatie steps could be taken due to the possibility of the spreading area having the high risk potential of dengue fever transmission.