In this paper, the neural network based modeling for the HfO2 thin film characteristics, such as the accumulation capacitance and the hysteresis index, grown by metal organic molecular beam epitaxy was investigated. In order to build the process model, the error back-propagation neural network was carried out and the X-ray diffraction data were used to analyze the characteristic variation for the different process conditions and predict the response models for the electrical characteristics. Principal component analysis was selected to reduce the dimension of the data sets. The compressed data were then used in the neural network. Those initial weights and biases are selected by Latin Hypercube Sampling method. This modeling methodology can allow us to optimize the process recipes and improve the manufacturability.