To accurately detect the damage for transverse guiding equipment of gear rack climbing type ship lift，considering the natural frequency change rate，stress，and strain as input characteristic parameters，we propose a structural damage detection model combining with the damaged structure classifier，damaged location classifier，and damaged level classifier. Taking Xiangjiaba ship lift as an example，we carry out the modal analysis and static analysis under 18 damage conditions to obtain 1，646 training samples and 100 testing samples. The structural damage detection model for the transverse guiding equipment was trained and tested based on the BP neural network，support vector machine （SVM） and Bayesian. The result demonstrated that the accuracy of the model based on the BP neural network for the damaged structure，damaged location，and damaged level are 93%，90%，and 91% respectively，which was 7% and 13% higher respectively than that based on SVM and Bayesian. The model was effective for the damage detection of the transverse guiding equipment.