Streaming is a method of distributing digital content directly over the internet, which allows users to access media without the need to download files. Bstation is a streaming platform that combines (OGV) and User-Generated Content (UGC). This research assesses the effectiveness of the K-Nearest Neighbors (KNN) and Naïve Bayes algorithms in analyzing sentiment in user reviews of the Bstation application, using a data sample of 5,000 reviews. The problem faced is the large number of users of the Bstation application, so sentiment analysis is needed to measure and understand the public's assessment of the application more accurately. This research aims to analyze the sentiment of Bstation users on Playstore and compare the performance of K-Nearest Neighbors (KNN) and Naïve Bayes to determine the best method for classifying reviews and user sentiment patterns. The findings showed that Naïve Bayes achieved 84% accuracy, surpassing KNN which only achieved 68%. Naïve Bayes showed 86% precision and 88% recall for negative sentiment, while achieving 78% precision and 76% recall for positive sentiment. recall for positive sentiment. In contrast, KNN achieved 80% precision and 66% recall for negative sentiments, and 54% recall for positive sentiments. recall for negative sentiments, and 54% precision and 71% recall for positive sentiments. The F1-Score for Naïve Bayes is also higher, reflecting a better balance overall. better balance overall. The macro average and weighted average weighted average for precision, recall, and F1-score with Naïve Bayes were 82% and 83%, respectively, while KNN recorded a macro average of 0.67. In conclusion, Naïve Bayes is more effective in sentiment analysis than KNN, providing more consistent and accurate performance