Ergonomic reliability plays a significant role in the safe operation of devices. With the spread of infectious diseases around the world, in work environments with high loads and high infection rates, medical staff work in a state of high self-protection. The use of visual display terminal (VDT) for medical equipment has undergone fundamental changes, and the traditional medical equipment human-machine interface design needs to be improved. After the completion of design and development, a VDT design enters the experimental testing stage, which has significant limitations for simulating the work of medical staff in the high-load and high-infection environments. The testing cost is high, and subjects face harsh conditions; thus, an ergonomic reliability model that can predict the use of VDT in such special high-infection and high-load circumstances must be established. An ergonomic reliability model based on an improved backpropagation neural network (BPNN) and human cognition reliability (HCR) is proposed for predicting and evaluating operation flows according to medical equipment VDTs. Firstly, a small data sample can be used to train BPNN to generate a network that can ensure suitable accuracy. To prevent the model from falling into local optimal solutions, the bat algorithm is introduced to improve the BPNN. Compared to a traditional BPNN, the superiority of the improved BPNN is clearly demonstrated. Secondly, the HCR method is used to analyze and highlight changes in the human factor reliability of VDTs for medical equipment in different time processes and operating processes according to BPNN prediction results, to provide a reference for selecting the optimal method. Finally, the validity and availability of the proposed method are verified through an eye tracker experiment and statistical analysis results.