Stakeholders widely use sentiment analysis in assessing sentiment towards an object. In this research, the object to be taken is sentiment analysis of political figures for the 2024 presidential candidate which is being widely discussed by netizens, especially on Twitter. The issues raised are regarding the performance measurement of an algorithm in classifying sentiments, some algorithms often need a higher level of accuracy. This study aims to improve performance measures from previous studies using the Naïve Bayes algorithm which has a fairly low level of accuracy, and in this study the SVM algorithm was used. This study takes Twitter data related to presidential candidates to see public opinion for each presidential candidate. The data taken was Twitter data with the keywords Ganjar, Anies, Prabowo totaling 8,959 data taken on October 17-25 2022. The results of the test concluded that the SVM algorithm has a performance measure or quite high accuracy compared to the Naïve Bayes algorithm in previous studies only of 73.86% while the SVM algorithm gets an average accuracy value of 98.61%, namely the Ganjar Pranowo dataset, then 98.81% precision, 99.79% recall. And for the proportion of sentiment, the positive sentiment obtained by Ganjar was higher than the other presidential candidates, namely 55%, Prabowo 30% and Anies 15%, while Anies' negative sentiment was 89% higher than Ganjar 8% and Prabowo 3%.