Selectivity estimation is a critical task for query optimizer. Producing an efficient execution plan depends on that estimation. To estimate the result size of the query that includes multiple attributes needs to approximate of the joint data distribution (JD) of the attributes. in addition, it is not enough to approximate this JD, but also it could be used effectively to estimate selectivity of the range predicates of the attributes. Most commercial DBMS use attribute value independence assumption (AVI) to calculate selectivity. But in real world such assumption almost is not true and causes large errors during cardinality estimation process. In this paper we propose two histogram-based approach to estimate multi-attribute range selectivity. In both techniques one-dimensional histograms are used, however, joint frequency matrix (JFM) is also constructed and used when it is allowed to improve accuracy of estimation. Simplicity and accuracy allow these techniques to be implemented easily and practically more useful.