Abstarct: Petrophysics and reservoir characterization is the main purposes of oil exploration. For Petrophysics and reservoir characteristics estimation with high confidence, to early reservoir simulations using seismic and well log data is essential. Using seismic attributes, hidden information extract from seismic raw data and finding relationships between seismic attributes and reservoir petrophysical parameters in place wells, to achive the distribution of these parameters with better accuracy where the well log data do not exist. The fundamental problem in the way of integration of seismic data and well data, is an integration method of this data with each other in order to estimate reservoir properties. Much research has been conducted in this area that make the estimation of petrophysical properties of the reservoir using seismic data (raw and indicators) and well log possible. The studies that have been conducted under the general tixtle of multi attributes analysis, often using conventional regression or neural networks methods to intedrate seismic and well log data and have reached significant results in descxription and estimation of the distribution of petrophysical parameters of the reservoir. Multi attributes classification is another method to integrate seismic and well logs data. Adaboost classification method which achive more attention in recent years and today it is one of the 10 superior data-mining algorithm. Adaboost outstanding feature is setting classification in the learning phase. The algorithm is fairly simple and fast solves overfitting pooblem there is in other methods and can also be combined with other techniques to form a fast classification algorithm. In this thesis the Adaboost method has used to multi attribute classification in a reflection seismology studies for classification of a horizon baxsed on petrophysical information and identify structures such as salt are used. Adaboost method has 78% accuracy in seismic horizon classification baxsed on porosity.Also it has 94% accuracy in the determining of salt dome boundaries from the surrounding media and the salt dome baxse which. Compared with SVM classification method that is 91 percent accurate, the Adaboost methods accuracy is higher.