The attitude adjustment of the wall-climbing robot is affected by its motion state,and it is difficult for the control center to correct the direction and angle of the wall-climbing robot in real time,and there is a problem of unbalanced motion of the robot.A control method for the attitude adjustment of a wall robot.The motion equation and statics equation of the robot were established to judge the unbalanced state of the robot.The RBF neural network was used to correct the posture of the wall-climbing robot,and the RBF neural network was used to control the posture adjustment and control of the wall-climbing robot.The experimental results show that after applying the proposed method,the average errors of yaw Angle,pitch Angle and roll Angle of the wall-climbing robot are 0.20×10³,0.15× 10³and 0.45×10³rad,respectively.The maximum errors of the reached position and the target position are 2 and 10m respectively,which reflects the excellent performance of the attitude control of the wall-climbing robot.