Considerable consideration should be given to the sentiment analysis of customer reviews when formulating a company's growth plan. As the internet has developed over the past decade, massive amounts of data have been produced across all professions. These developments have given people new platforms for sharing their thoughts on things, such as Google Reviews, Blog Posts, Tweets, etc. Sentiment analysis is the technique of statistically recognizing and categorizing emotions stated in a piece of text, specifically to discover whether the writer's perspective towards a particular thing is positive, or negative. Social media provides a robust platform for gathering and analysing customer input, which can be leveraged to expand business opportunities and better serve customers. Therefore, the reviews left by customers will be examined closely in this study. Kaggle data is used to complete this study. The data includes comments about the restaurant and the reviewer's attitude towards the restaurant. The customer review is in string format. A customer's positive opinion of the restaurant is represented by the number 1, while a negative opinion is shown by the number 0. Certain processing methods, such as the elimination of irrelevant information, are required for the review's string data. Later, the Machine Learning (ML) model will receive the cleansed data. As an ML model, we've settled on the Support Vector Machine (SVM). To decipher how customers, feel about a certain restaurant, a basic SVM and an optimized SVM with Particle Swarm Optimization (PSO) are used. From the finding, we found that PSO-SVM architecture achieves greater accuracy of 85.5% accuracy. We also measured our model's efficacy using a variety of alternative metrics. When compared to SVM, PSO-SVM will perform better on all metrics.