Concrete is made from cement, water, aggregates, additives and mineral additions. In its fresh state, one of its properties is the settlement. Meanwhile, when hardened, its mechanical properties are reached at an age of 28 days after the mixing of the raw materials, as a brittle material it requires fiber to be incorporated to acquire ductility. Both in concrete, without fiber and with fiber, the use of artificial neural networks (ANNs) to predict mainly compressive strength has been expanded, which, as well as for its settlement properties, it also makes its other properties interesting . In this paper, the development of ANN is reported in combination withtraining from Levenberg-Maquardt and Scaled Conjugated Gradient Algorithms, using MATLAB software to predict the settlement and design strengths of compression, tension, shear, and flexural strength, flexural toughness, and for steel fiber reinforced concrete. The results of correlation between actual and predicted values show that the computational tool developed is reliable for predictive use. Keywords: Concrete´s settlement, Fiber reinforced concrete, Artificial Intelligence, Predictive tool, Concrete’s mechanical properties, Artificial Neural Network.