Sentiment analysis could be a powerful tool in evolving psychotherapy. There is a rapid increase in patients seeking mental health help and NLP could help make their experience more accessible and efficient. A sentiment analysis-based journal could help users track their thought patterns, their severity, and progress over time. This paper investigates the effectiveness of Naïve Bayes, Random Forest, Support Vector Machine, XGBoost and BERT algorithms in emotion classification of text for future journal use. Comparative analysis helps understand which algorithms could be best suited for this type of multi-label classification, and broadens current research by testing several algorithms which can show what should be further worked upon in the field, and which algorithms are best to avoid. A lot of studies test only one or a fewer number of algorithms, leaving less room for comparison on the same dataset under same conditions, so it is unclear if accuracy differences in different studies are derived from a better model or a better dataset. The four algorithms were trained on a dataset of 17.449 emotion-annotated sentences after preprocessing steps including tokenization, lemmatization, and TF-IDF vectorization for feature extraction. Naïve Bayes performed the worst with a 71% accuracy and 0.66 F1-score, while other algorithms all reached 87% accuracy and a 0.87 F1-score, which shows significant improvement to other research in the field. Furthermore, BERT accomplished 93% accuracy, making it the best performing model in the study. These findings demonstrate the potential of sentiment analysis in aiding psychotherapy needs, whether it be therapists, patients, or users keeping track of their mental health.