The accuracy of growing stock estimates depends on the availability of local volume tables to infer volume of trees from forest inventory data. The current study was undertaken to prepare local volume tables for Cedrus deodara of Gilgit-Baltistan (GB). Data was collected from 52 sample trees which were measured for the study in natural dry temperate forests of Gilgit-Baltistan. Diameter at Breast Height (DBH) and total height of the sample tree were measured before felling. After felling, the bole was cut into 2m logs with end log of variable length. The over bark mid diameter of the log and its length were measured for determining volume of logs. Total volume of a tree was determined by adding volumes of all logs. Different regression models were tested for deterring best relationship between DBH and height and DBH and Volume. On the basis of the best fit models, volume tables were prepared in 2cm dbh classes. These tables yielded volume estimates closer to the actual values obtained field measurement.