One of the important issues of Magnetorheological (MR) elastomer characteristic is to determine the linear viscoelastic region (LVE). This lead to the number of prediction models to predict the characteristic of MR elastomer based on its rheological properties. Based on the previous research, it is found that the availability of prediction model is regarded as one of the concerned issue in the development of MR elastomer material and devices. Therefore in this work, Artificial Neural Network (ANN) is proposed as prediction model due to its advantages particularly in terms of non-parametric modeling approach flexibility. Here, the shear strain and magnetic flux density are adopted as the inputs of the prediction model. Meanwhile the predicted outputs were the storage and loss modulus. The prediction model performance index was analyzed based on its generalization performance. Comparison between simulation result and experiment data shows good agreement where prediction on 0.80 T of output storage modulus produced 0.0086 MPa and 0.998 for Root Mean Square Error (RMSE) and Coefficient of Determination (R 2) respectively.