Abstarct: Identification of undesirable conditions in road pavement is considered as an important factor in road pavement management . for this purpose , various models such as probabilistic parser and regression analysis are used to determine the pavement quality . in order to predict pavement conditions in the future it is necessary to use indicators that one of them is mechanical properties of asphalt . in this end , marshall strength, marshall flow and resilient modulus are modeled as a function of asphalt mix design specifications . using the data of the comprehensive databaxse of aggregate performance information of asphalt , three models have been proposed using three main parameters including marshall strength, marshall flow and resilient modulus as dependent variables and the percentage of bitumen binder content , the largest binder size of aggregates , the percentage of asphalt penetration at 77 degrees fahrenheit , and void percentage between aggregates as independent variables . this is a type of regression model . to do so , raw data including all of the above parameters were extracted from LTPP databaxse . then , using spss statistical analysis software . using these data , eight different models were suggested for each of the main asphalt variables and the model that had better performance was selected . the selected model with correlation coefficient in range of 0.7 to 0.9 has shown good accuracy due to the distribution of data .