Project can be classified to small, medium, large and very large. This could be mainly based on number of activities, the budget of the project or the completion time. The criterion of scaling could be change from country to another. In this project, it is assumed that as the number of activities increase, the scale is increased. This thesis develops an algorithm to generate a unique Activity On Arrow (AOA) network with a minimum number of dummy activities. It also develops an algorithm to deal with multi resources, by generating a new resource, which is in relation to all considered resources by a defined expression. A Genetic Algorithm is used in order to perform resource leveling and allocation in the large-scale project. The developed algorithm is based on shifting the noncritical activities within their total float to reduce the undesired fluctuation. Four indices are considered to determine the optimum scheduling process. A mathematical model is developed to consider these indices by creating a weight for each one; a simulation is applied in developed MATLAB program to reach to the best weights that achieve the optimum scheduling. The efficiency of the proposed algorithm is measured by the resource improvement coefficient. As a case-study, an expansion took place of the Irbid Specialized Hospital project, in Irbid-Jordan, is used to explore the application of the proposed algorithm. For this real case, the proposed algorithm gives almost normal distribution shape of the activity leveling, which validates the algorithm applicability.