Abstract: The balise is an important equipment in the signal control system of urban rail transit, which plays an important role in the accuracy of train parking and the transmission of mobile authorization under the control level of point trains. This paper analyzed the layout structure and properties of equipment of urban rail signal system, and used CAD secondary development technology and based on the. Net platform to provide an automatic generation method for precise installation data of balise, which improved the efficiency of data configuration. In response to the issue of ensuring data accuracy, the paper proposed a model based on Deep Neural Network (DNN) for automatic verification of precise installation data of the balise, and took Nanchang Metro Line 3 as an example to experimentally validate the model. The results show that the automatic generation algorithm for precise installation data of the balise can accurately generate data. Using a three-layer DNN model and ADAM (Adaptive Moment Estimation) optimizer, the accuracy of the validation model can reach 95.45% when the Batch Size is 256 and the training and validation set segmentation ratio is 7∶3.