In this paper, inverse fuzzy models obtained online for uncalibrated visual servoing, are developed and validated in a six degrees of freedom robotic manipulator. This approach will recursively update the inverse fuzzy model based only on measurements at a given time instant. The uncalibrated approach does not require calibrated kinematic and camera models, as needed in classical visual servoing to obtain the Jacobian. Experimental results obtained in a PUMA robot performing eye-to-hand visual servoing are used to demonstrate the validity of the proposed approach, when compared to the previous developed off-line learning.