With the advantages of low fuel consumption, high departure efficiency, and green environmental protection, the new civil aircraft departure mode of towing taxi-out has become the preferred mode of civil aircraft departure for future smart airports. The process of civil aircraft towing departure requires close collaboration and effective communication among multiple staff members, and accidents are mostly caused by human errors. Therefore, it is crucial to perform human reliability analysis. However, there are few reports or human reliability studies on civil aircraft towing departure accidents. Although the Success Likelihood Index Method (SLIM) based on expert ex-perience may be an effective tool for assessing human reliability, SLIM relies heavily on expert judgment to determine model parameters, which makes the method subjective and uncertain to varying degrees. To overcome this, the present study uses the Demp-ster–Shafer evidence theory to extend SLIM by fusing the judgments given by different experts, providing more accurate and reliable results. Herein, the proposed model is used to study human reliability in the process of civil aircraft towing departure. The results are conducive to reducing the human error probability of the staff involved in the civil aircraft towing departure process and improving the overall safety level of departure operations.