This article explores modern methods of predictive analytics used to assess short-term and seasonal demand fluctuations in retail trade. It analyzes the effectiveness of time series models, machine learning algorithms, and neural network approaches in a dynamic market environment. Particular attention is given to data preparation, the selection of accuracy metrics, and the integration of forecasting models into logistics and inventory management. Hybrid solutions are emphasized as crucial, and tuning models to outside drivers such as weather conditions and promotional activity are emphasized. The study concludes that predictive analytics is highly practically relevant to enhance the resilience and effectiveness of retail supply chains.