Hotel reviews are obtained from consumers’ experiences after using hotel accommodation services as textual information. This information can help potential customers to choose services that suit their needs. Prospective customers are pretty difficult to judge the hotel from the overall reviews. To make consumers more accessible to determine hotels, this study builds a sentiment analysis model that can distinguish between positive and negative sentiments. This study created a sentiment analysis model using the Word2Vec extraction feature with the KNN classification algorithm. The final result in this study produces the best model using a stemming dataset. The vector dimensions of the Skip-gram model architecture are 300, and the use of K=3 value in the KNN algorithm. The model can produce an accuracy value of 82.61% and an f1 score of 83.96%.