Abstarct: Drilling operations is one of the most costly activities in the oil industry upstream. Drilling industry has focused on reducing costs. The time factor can be used in drilling operations as the most effective factor of rising costs. Most of the costs associated with drilling, including personnel costs, drilling rig rental, transport and feeding, directly increase by increasing time of a well drilling. Losing of drill bit into the bore hole, reducing the penetration rate, fishing, redirect the well due to loss of the original route or direction of trajectory and stuck piping due to pressure difference can be named delay in drilling operations. In this study, we try to introduce model against traditional methods of wearing bit control conditions that require direct observation and stop drilling. So engineers be able to control the bit wear condition real-time in drilling of the well bore. Thus, by preventing of rate reduction of drilling penetration rate due to reduced wear bit and cutting power for drilling the formation, it will reduce the cost of the drilling operation. By taking advantage of researches and previous studies had done in this field on the basis of mathematical models and neural networks, we will try to be a step further. New models are baxsed on adaptive neural fuzzy inference systems and other model is neural network model using a genetic optimization algorithm, instead of gradient baxse algorithm method, in order to compare with fuzzy logic and neural networks and to improve the construction of a model to predict wear of the bit. Model input parameters are included: the drilling penetration rate, weight on bit, rotational speed, pump flow, depth in and depth out, bit size, bit type and time. Degree of the bit wearing is the output. According to the correlation, coefficient value of test data is 0.96444 for the first model, and 0.92524 for secend model. result was close to observe value. This models are also reliable for the perediction.