Aiming at the delay of video transmission in the working process of teleoperation hydraulic engineering robot, a visual cue method based on attitude recognition for teleoperation hydraulic engineering robot was proposed. The camera was used to acquire the image of the teleoperated hydraulic engineering robot, and the keypoints identification stacked hourglass (KPR-SHN) model and the attitude recognition neural network (AR-NN) model were used to obtain the displacement of the hydraulic cylinder of the teleoperation hydraulic engineering robot, thereby driving the virtual engineering robot implement a visual cue to the operator. The method has the characteristics of simple structure and high technical robustness.Experimental results show that the method is simple and effective, and the error is less than 8%, which can provide guidance for the design of visual cue system based on virtual reality for teleoperation engineering robots.