Production and distribution scheduling are quite important from both theoretical and practical points of view. Integration of production and outbound distribution scheduling has recently attracted significant amount of research. In classical scheduling problems, coordination with traortation unit and dispatching situations has not been involved directly and decisions about them have been made independently. On the other hand integrated decisions on product scheduling and distribution planning, which is a more comprehensive view to this problem, decreases the costs of production system, lead to more profit, improve quality of service and increase the satisfaction of customers. Althoughbecause of being many applied problems, the researches on this area have been increased in recent decades, there is not a significant study on “integrated decisions on production scheduling and distribution planning”. Due to importance of “objective function of tardy cost” (total weighted number of tardy jobs) on the satisfaction of customers, in this study we tried to propose exact and inexact methods to solve the problem of “scheduling of similar parallel machines” with “delivery batching” using “minimum tardy cost and delivery cost objective function”. In this study two models of mixed integer programming are developed to solve the mentioned problem. After this, for presenting a more powerful exact method for the considered problem, We suggested a branch and bound algorithm. To implement our approach, we developed a heuristic method to create a primary upper bound, lower bounds and instructions which decrease the answer space. The proposed branch and bounding method can solve 86% of problems with 20 tasks and 5 machines in 3600 seconds. Since in practice there are many large scale problems which cannot be solved in a reasonable time, a Genetic Algorithm approach is also proposed to approximate the true answers in a shorter time for large scale problem