Abstract Unconventional reservoirs are characterized by their extremely low permeabilities surrounded by huge in-situ stresses. Hydraulic fracturing is a most commonly used stimulation technique to produce from such reservoirs. Due to high in situ stresses, breakdown pressure of the rock can be too difficult to achieve despite of reaching maximum pumping capacity. In this study, a new model is proposed to predict the breakdown pressures of the rock. An extensive experimental study was carried out on different cylindrical specimens and the hydraulic fracturing stimulation was performed with different fracturing fluids. Stimulation was carried out to record the rock breakdown pressure. Different types of fracturing fluids such as slick water, linear gel, cross-linked gels, guar gum, and heavy oil were tested. The experiments were carried out on different types of rock samples such as shales, sandstone, and tight carbonates. An extensive rock mechanical study was conducted to measure the elastic and failure parameters of the rock samples tested. An artificial neural network was used to correlate the breakdown pressure of the rock as a function of fracturing fluids, experimental conditions, and rock properties. Fracturing fluid properties included injection rate and fluid viscosity. Rock properties included were tensile strength, unconfined compressive strength, Young's Modulus, Poisson's ratio, porosity, permeability, and bulk density. In the process of data training, we analyzed and optimized the parameters of the neural network, including activation function, number of hidden layers, number of neurons in each layer, training times, data set division, and obtained the optimal model suitable for prediction of breakdown pressure. With the optimal setting of the neural network, we were successfully able to predict the breakdown pressure of the unconventional formation with an accuracy of 95%. The proposed method can greatly reduce the prediction cost of rock breakdown pressure before the fracturing operation of new wells and provides an optional method for the evaluation of tight oil reservoirs.