This study presents a financial distress prediction (FDP) approach for Vietnamese-listed firms, utilizing 19 financial ratios and a sentiment analysis feature extracted from news articles. This marks the first research effort on the Vietnamese market to incorporate text-based features for improving FDP performance. Sentiment scores were computed by applying a pre-trained BERT model (vinai/phobert-base) to news data, followed by calculating the average annual sentiment score for each firm. The dataset comprises 12,678 firm-year observations, with the Synthetic Minority Oversampling Technique (SMOTE) applied to the training set to address class imbalance. Experimental results demonstrate that the Random Forest outperforms other machine learning algorithms, achieving an F1-score of 85.90% and a recall of 84.19% on the test set. The integration of sentiment analysis scores notably enhances predictive performance, with F1-scores averaging an improvement of 1.72% compared to models using only financial ratios. Feature importance analysis further highlights the roles of both sentiment scores and key financial indicators, such as profitability, emphasizing their complementary contributions to the predictive framework in Vietnam’s emerging market.