From the perspective of rapid global development, inventory planning and management face many challenges, including order fluctuations and demand uncertainty. To this end, enterprises should significantly adjust inventory turnover to improve and reduce costs, thereby formulating a more robust inventory policy. Therefore, this paper will analyze the application of artificial intelligence (AI) in cross-border e-commerce inventory control and forecasting to improve the efficiency of its supply chain. This paper collects historical sales data from the past three years and takes into account characteristics such as daily sales volume, promotional activities, seasonal effects, and market trends. The dataset is cleaned and preprocessed through different data cleaning and preprocessing techniques to handle missing values, outliers, and feature engineering is performed to score important features. The data is then put into the Long Short-Term Memory Network (LSTM) model for fine-tuning to achieve demand forecasting. The prediction accuracy of the LSTM model is as high as 95.1%, and the average inventory turnover rate can reach 75.4%, thereby reducing inventory backlogs and out-of-stock phenomena. Artificial intelligence technology intervenes in all aspects of cross-border e-commerce inventory management, with scientific data guidance, to achieve the maximum optimization of enterprise inventory decisions and improve overall operational efficiency.