In enhanced oil recovery operation, well placements, well depths, injection, production rates and time and the number of injection and production wells are opted as decided parameters. It is considered in order to finally achieved maximum amount of possible oil may be extracted with the least charge in certain time. This matter makes be declined water loss and increased more oil production. Because of finding optimum point of production, a suitable simulation from oil reservoirs may be an important subject. In optimization process, by using numerical simulation with a set of coupled partial differential equations may be extremely time consuming and expense in order to find the optimum parameters. Moreover, for simulation of oil reservoirs and providing required fitness function in optimization, artificial neural networks have been used. Neural networks were made with iired by the nervous system of human body that to be suitably trained with numerical simulation data in order to be used as an approximation of fitness function (a function to be found its optimum point). The optimization method with iired by nature was invoked. Entitled genetic algorithm as an implementation which is a stochastic method and extremely capable to find the global optimum point to escape from local traps (local optimum point). So, this method is very useful, with high certainty and also definitely sufficient which helps to be achieved the global optimum point. In this thesis, the effects of different reservoir parameters on the optimization of oil reservoirs have been considered. Effective parameters in optimization such as the number and placements of injection and production wells and also injection rate were regarded. In addition, effective parameters on genetic algorithm such as the number of initial population, mutation and crossover rate have been investigated. This method was shown its capabilities to find global optimum point quickly. It may be used in applicable programs and the simulation of oil reservoir due to its flexibilities relative to change reservoir status to optimize oil reservoirs in real time due to efficiency. Also, the results have been demonstrated to increase the number of production wells in production trend from reservoir with no positive effects and sometimes, it made production from reservoirs get instructed. Of course, to increase the number of injection rate, it should be noticed that with increasing the number of them relatively caused to go up production from reservoir. Despite with increasing it, elicitation may be decreased from reservoir. It should be noticed related to increasing the number of injection wells which may be made increase elicitation from production wells however with increasing it more cause to decrease oil production. Case studies have been implied that using more injection rate at first and increasing injection operation then, should significantly help in production performance from wells. Keywords: Oil reservoir simulation, Artificial Neural Network, Finite Volume method, Genetic Algorithm, Well placement