This study investigates the effectiveness of deep learning methods for sentiment analysis of financial tweets in Turkish. The researchers collected financial tweets related to the BIST100 index and categorized them as positive, negative, or neutral. They created binary and multi-class datasets and used word embedding and pre-trained word embedding for tweet representation. Various deep learning models, including Neural Network, CNN, LSTM, GRU, and GRU-CNN, were tested for sentiment classification. The best results were achieved with the CNN model using pre-trained word embedding, achieving 83.02% accuracy for binary classification and 72.73% for multi-class classification. The GRU-CNN model performed best for binary classification (80.56%), while the Neural Network model performed best for multi-class classification (63.85%). The study highlights the importance of using deep learning methods for sentiment analysis in financial contexts, particularly for Turkish financial tweets. The results suggest that pre-trained word embeddings and CNN models are effective for sentiment classification in financial data. The study also acknowledges limitations, including the relatively small dataset size and potential biases in data collection and labeling. The findings have implications for financial decision-making, as sentiment analysis of financial tweets can provide valuable insights into market trends and investor sentiment. The study contributes to the field of financial sentiment analysis by demonstrating the effectiveness of deep learning methods in Turkish financial data.This study investigates the effectiveness of deep learning methods for sentiment analysis of financial tweets in Turkish. The researchers collected financial tweets related to the BIST100 index and categorized them as positive, negative, or neutral. They created binary and multi-class datasets and used word embedding and pre-trained word embedding for tweet representation. Various deep learning models, including Neural Network, CNN, LSTM, GRU, and GRU-CNN, were tested for sentiment classification. The best results were achieved with the CNN model using pre-trained word embedding, achieving 83.02% accuracy for binary classification and 72.73% for multi-class classification. The GRU-CNN model performed best for binary classification (80.56%), while the Neural Network model performed best for multi-class classification (63.85%). The study highlights the importance of using deep learning methods for sentiment analysis in financial contexts, particularly for Turkish financial tweets. The results suggest that pre-trained word embeddings and CNN models are effective for sentiment classification in financial data. The study also acknowledges limitations, including the relatively small dataset size and potential biases in data collection and labeling. The findings have implications for financial decision-making, as sentiment analysis of financial tweets can provide valuable insights into market trends and investor sentiment. The study contributes to the field of financial sentiment analysis by demonstrating the effectiveness of deep learning methods in Turkish financial data.