The Vaovao Malagasy Sentiment Corpus (VMSC) is the first publicly available, manually annotated sentiment analysis dataset for the Malagasy language (mg). It contains 5,041 sentences extracted from news articles (vaovao) published between 2022 and 2023.