Abstarct- The task of named-entity recognation (NER) can support the higher-level tasks such as question answering, text summarization, and information retrieval. This work views NER on indonesian Twitter posts as a sequance labeling problem using supervised machine learning approach. The architecture used is long shirt -term memory networks (LSTMs), with word embedding and POS tag as the model features. As the result, our model can give a perfomance with an F1 score of 77.08%