Drought as an unavoidable phenomenon and natural catastrophe has been occurring in different zones of countries especially those located in arid and semiarid regions frequently. In recent years this phenomenon has been becoming more acute and caused too many causalities and damages. Drought predictions serve as the most important combat and reduce drought damages. Recently to calculate and evaluate streamflow drought index(SDI) has been received too many attention as one of the hydrological drought prediction approach. In frequency analysis of this index, we are encounterinto short term statistical period and basins without measurement stations which to overcome them, the frequency analysis regionalization method is used. On the other hand since droughts occur as regional in vast area for long term periods, to study them in regional concept is important. In the present research we exploited frequencyregional analysisof streamflow drought indexto predict hydrological drought in northKaroon basin. For this, weselected the number of fourteen appropriate stations and collected their statistics. Since one of these stations had less recorded data and due to short statistical period, we eliminated this stationand conducted regional analysis with other ones. To investigate homogeneity of study area, weconsidered and climatic physiographical features of elite stations and to carryout homogeneity test we used Husking and Wallis (1991) and cluster analysis. Results of cluster analysis showedthat study area might be consideredas ahydrological homogenous region. Then in respect to availablestations discharges and SDI calculation formula we calculated this index value in time series of 3, 6, 9 and 12 months. Basedon Chi square goodness of fit test and kolomogrophSmirnoff three parametric distributions was identified as a best regional distribution. Accordingly,SDI index of 3,6,9 and 12 months with return periods of 2,5,10,50 and 100 years using FERQ program in every station. To conduct regionalfrequency analysis and relationship between differentparameters and SDI values with different return periods, multivariate regression was used and as a whole, the number of thirteen climatic and physiographic features was applied to prepare regional models. Results of regional analysis indicatedthat area, concentration time, main waterway length, main waterway slope, Miller coefficient and Gravelious one are the effective factors on study area SDI index.In terms of obtained results, the region was justify; MARGIN: 0cm 0cm 0pt" Keywords :drought, hydrological drought prediction, surface flow drought index,(SDI), northernKarron, regional analysis