The flower market is highly sensitive to seasonal trends, supply chain disruptions, and demand volatility, making accurate forecasting essential for financial sustainability. Despite the availability of data, many businesses in the sector struggle to make informed decisions due to unpredictable market behavior. This research introduces a predictive analytics-based framework named Dynamic Vortex Search-driven Intelligent Long Short-Term Memory (DVS-ILSTM) to support data-driven decision-making. It gathers historical sales data, market trends, and environmental indicators from publicly available sources and internal company records. For pre-processing, the Min-Max Scaling technique is used to normalize the data, ensuring all variables are on the same scale for better model performance. Feature extraction is achieved through Independent Component Analysis (ICA), allowing critical patterns to emerge. The DVS-ILSTM combines Vortex Search optimization to tune hyper parameters dynamically and the ILSTM model to capture temporal dependencies for accurate demand forecasting. This hybrid model supports decisions on production, inventory, and distribution with higher precision. By integrating predictive analytics into operational planning, businesses manage resources, reduce waste, and enhance profitability. Experimental results show improved forecasting performance across RMSE and MAE metrics compared to conventional models, as well as a higher R² value (67%). The proposed approach strengthens strategic planning and contributes directly to the financial sustainability of businesses operating in the competitive flower market