Emotion classification from online content has received considerable attention from researchers in recent times. Most of the work in this direction has been carried out on classifying emotions from informal text, such as chat, sms, tweets and other social media content. However, less attention is given to emotion classification from formal text, such as poetry. In this work, we propose an emotion classification system from poetry text using a deep neural network model. For this purpose, the BiLSTM model is implemented on a benchmark poetry dataset. This is capable of classifying poetry into different emotion types, such as love, anger, alone, suicide and surprise. The efficiency of the proposed model is compared with different baseline methods, including machine learning and deep learning models.