This study analyzes the applicability of artificial intelligence (AI) in forecasting the Vietnamese stock market by integrating financial indicators with investor sentiment. The research methodology encompasses a literature review, secondary data analysis, and cross-referencing of official domestic and international sources. The findings indicate that price and trading volume data do not fully reflect investor expectations, attention, and reactions to new information. Consequently, the study proposes an AI data framework structured sequentially: macroeconomics - financial indicators - market information - attention - sentiment - trading behavior - price movements. Random Forest and XGBoost are identified as suitable for tabular data; whereas LSTM and Transformer architectures can effectively process timeseries and textual data. The research underscores that data quality, temporal synchronization, interpretability, and data leakage control are critical determinants of forecasting efficacy. The results establish a foundation for developing a multi-source AI forecasting system tailored to the distinct characteristics of the Vietnamese stock market.

Nguyen Quoc Thinh,"AI-Based Stock Market Forecasting: Integrating Financial Indicators, Investor Sentiment, and Attention" International Journal of Advanced Multidisciplinary Research and Educational Development, V2(5): Page(377-385) September - October 2026. ISSN: 3107-6513. www.ijamred.com. Published by Scientific and Academic Research Publishing.