China’s economy has developed rapidly, but its ecological civilization has been seriously affected. Traditional theoretical e-commerce construction cannot improve the efficiency of e-commerce construction. Therefore, it is of great practical significance to conduct in-depth research on countermeasures for e-commerce construction from the perspective of artificial intelligence algorithms. This article analyzes the problems in the construction of EC (which may refer to related systems such as e-commerce), and uses BP neural network to train the input and output layers, which can quickly identify environmental hazards, detect ecological hazards, and improve the quality of EC; Also utilizing it to identify and classify garbage, incorporating the network into input data learning and outputting categories; Implement non-linear mapping of input and output through BP neural network, transmit signals and predict future natural disasters. In addition, in addition to technical analysis, other countermeasures are proposed for EC construction issues in order to provide inspiration for e-commerce construction. After analyzing the BP neural network, this article conducts experimental comparisons on the recognition and classification accuracy of different types of garbage in e-commerce. The results showed that the BP neural network had recognition accuracy and classification accuracy of over 80% for recyclable waste, kitchen waste, and hazardous waste, with specific rates of 81.21%, 85.00%, 83.70%, and 81.51%, 83.29%, and 87.34%, respectively. The data proves that BP neural network is powerful in garbage classification and recognition, contributing to the construction of EC.