In order to improve the automation level of high-altitude operation of a transmission tower and ensure the safety of high-altitude operation robot equipment, aiming at the problem that traditional RRT and A* algorithms cannot quickly plan the safe trajectory of a manipulator in the high-altitude complex environment of an angle steel tower, a two-way beetle informed rapid random expansion tree trajectory planning algorithm (BI-RRT* Connect) based on the tower model is proposed. Firstly, the working environment of the high-altitude robot is reconstructed digitally, and the accurate model information of the transmission tower is obtained by using the model registration method. Then, according to the tower model, the feasible path between the end point of the manipulator and the working target is sought, and the beetle antennae algorithm is added to improve the random sampling process. Two beetles are used to find each other to quickly obtain the initial path solution. Finally, the initial path is used as a priori knowledge to construct the optimal path of the elliptical sampling area. The experimental platform is built in the Unity environment to compare the proposed algorithm with four traditional algorithms such as RRT* Connect and A*. The experimental results show that the algorithm has fast solution speed and high solution quality in various operation areas of the transmission tower.