Aiming at the problems of low accuracy， slow convergence and local optimality of traditional dung beetle algorithm in 3D unmanned aerial vehicle （UAV） path planning， a multi-strategy improved dung beetle optimizer （MSIDBO） is proposed. Firstly， spatial pyramid matching（SPM） chaotic mapping and reverse learning strategy are used to initialize the population to improve the diversity and quality of the initial population. Secondly， an improved boundary convergence factor is introduced to achieve the balance between global exploration and local search. Then， the attack mechanism of gull optimization algorithm is integrated to improve the convergence speed and solving accuracy. Finally， the t-distribution differential variation strategy is used to improve the ability of the algorithm to jump out of the local optimal solution. The improved Dung Beetle algorithm is compared with other heuristic algorithms and related improved algorithms by benchmark function test. Compared with other heuristic algorithms and improved algorithms， MSIDBO algorithm has outstanding performance in convergence speed and accuracy. In addition， the improved Dung Beetle algorithm is applied to 3D UAV path planning simulation. Experimental simulation results show that the path cost function generated by MSIDBO algorithm is smaller， the path quality is higher， and the stability is better under different scenarios.