This paper presents a method based on neural network (NN) for estimating the properties of semiconductor thin film. Through the effective learning process, NN is able to catch the relationship between input and output pairs bypassing the complicated statistical steps such as model hypothesis, identification, estimation of model parameters, and verification. Such an estimator then can be developed to be a smart mechanism which can help the technician to set the relevant control parameters in the manufacturing process of thin film. In this research, the thickness and refractive index (RI) of thin film were estimated by the well learned NN model. From the studied results shown, the properties of thin film indeed could be estimated in advance according to the relevant control parameters in the manufacturing process. That also means the estimator we developed could be built and fulfilled its function.