We describe here a method for estimating the pouring flow rate for tilting-ladle-type automatic pouring robotused in casting industries. To precisely pour molten metal into the mold, controlling the flow rate of liquid flowing outof the ladle is mandatory. However, it is difficult to directly measure the pouring flow rate by using a conventional flowmeter, because the flow meter is damaged by the high temperature molten metal. Therefore, in this study, we used a softsensing technique as part of the pouring flow rate estimation system. The extended Kalman filter was applied to the flowrate estimation of the automatic pouring robot in previous study. The previous approach can be applied only the ladlewith smooth shape, because of derivation of Jacobian matrix. However, some ladles with complicated shape which theJacobian matrix cannot be derived have been used in an actual casting plant. In this paper, an unscented Kalman filterwhich does not need the Jacobian matrix is applied into the flow rate estimation. The effectiveness of the proposed flowrate estimation method is demonstrated through simulations and experiments.