Due to the nonlinear characteristics between the input & output parameters of a liquid flow rate process control, classical optimization technique is limited for this purpose. Hence computational optimization is chosen as an alternative approach. In this paper an ANFIS model was designed using trial and error based on various three different sets of experimental data sets for checking the flexibility, speed & adaptability of these soft computing technique. By the understanding of the Sugeno type ANFIS structure parameters are set to facilitate the hybrid learning rules. However, it is seen that by increasing the number of inputs response time of the model also increased. The results are in good agreement with the experimental results & can be applied to predict the performance of mass flow sensor. For the best ANFIS structure gained in this study RMSE and MAE were calculated as 2.143 & 0.504 respectively.