Landslide is a kind of natural disasters that annually causes great losses in life and property throughout the world. With the increasing population and expanding urban areas to the steep slope areas the risk of landslide has increased. The objective of this research is to study the risk of landslide in order to reduce potential damage and manage landslide risk. Siminehrood basin is the case study of the research. It is a subbasin of Orumiyeh lake catchement area at north west of Iran and due to its geographical location and natural characteristics is one of the country's landslide prone areas.

In this research nine main factors affecting landslide including distance to fault, distance to road distance to stream network, topography, slope, aspect, landuse, precipitation and lithology are used. The weights of evidence model which is a bivariate statistical analysis and data-driven technique is used for producing Landslide susceptibility mapping. For obtaining the weights of  evidence  modeling  prior probability,  conditional  probability  and  the  positive  and  negative  weights  are calculated and a layer describing the total evidence weight for each class of factor layer is produced. Prior to combining the weights of factor layers, the importance weight of each factor layer is calculated using the Analytic Hierarchy Process (AHP).

For implementation of the research the related spatial data were gathered and prepared in order to produce spatial layers for describing effective factors. For preparation of spatial data different processing have been performed such as distance tool for creating distance layer, interpolation tool for creation surface layer from point layer, hydrology tool set for extracting stream network from Digital Elevation Model (DEM) and classification tool. All of the preparation processes, calculation of importance of factor layers by using AHP method and overlaying of factor layers have been implemented in ArcGIS software package. A python application has been developed in ArcGIS environment in order to calculate the importance weight of each factor layer by using AHP. Finally, the landslide susceptibility map of Siminehrood Cachment was produced with overlaying the total evidence weight layers and applying the importance weight of each layer and then the resultant map was classified.

For assessing the results of research the total occurred landslide data set in study area was randomly split into two training and testing landslide data sets. Calculation of success rate of landslide susceptibility map using testing landslide data sets indicated that about 60% of landslide pixels which has been used for producing classified landslide susceptibility map are in high risk class and the density of landslide pixels in this class is 37% while in the medium risk class is 8.8% and in the low risk class is 0.083%. Predicting rate of resultant map shows that about 56% of landslides which have been used for verifying and assessing the research results are in the high risk class of classified landslide susceptibility map.