<p>Landslide inventory plays a vital role in landslide susceptibility zonation mapping. The landslide inventory maps may either be represented as a polygon or as a point. The present study deals with the comprehensive study on the influence of polygon and point landslide inventory on landslide susceptibility zonation (LSZ) mapping. For this purpose different landslide susceptibility zonation maps were prepared for a landslide-prone district in the Uttarakhand state (Indian Himalayas) based on 14 thematic layers along with a landslide inventory of 352 landslides. Two landslide inventory maps were prepared which contains two different spatial representations (polygon and point). For the generation of LSZ maps, two statistical models, namely frequency ratio (FR) and certainty factor (CF) were employed. The effect of landslide inventories (polygon and point) were examined based on four approaches (i) accuracy based on area under the ROC curve (ii) difference Image analysis (iii) assessment of LSZ maps with 100% polygon and point landslide inventories (iv) exchange of testing inventories for validation. It is found that the FR model outperforms the CF model for both the point and polygon inventories. The LSZ maps prepared using point inventory both for FR and CF models have higher accuracy than the LSZ prepared using polygon inventory. However, the validation of point testing inventory over the developed LSZ model using polygon inventory displays a better accuracy. On the other hand, the polygon inventory serves better to identify the landslide susceptibility zones with high clarity in case of landslides falling in more than one zone. This study will be helpful for the selection and preparation of landslides inventories that will enhance the accuracy and acceptability of landslide susceptibility maps.</p>