Representing the traveling path with origination and destination is challenging and crucial for Transportation Recommendation. Transportation recommendation has a goal of recommending a travel plan which considers various transportation modes, such as walking, cycling, automobile, and public transit, and how to connect among these modes. Previous methods employ multi-modal transportation graph[1] and some POIs to represent the traveling path. Although these methods achieve satisfactory results, they do not consider the ordered sequence which is always realistic in the daily traveling in people's life. For this purpose, we introduce a Bidirectional Encoder Representations from Transformers[2] for representing Transportation (BERT4Trans). We considered the continuity of ordered traveling plan and trained the sequence of path from origination to destination Inspired by the word2vec[3] and item2vec[4]. Our method has been used in our solutions of Transportation Recommendation and got a significant improvement in the weighted-F1. The code is in the https://github.com/chenweilong915/Bert4Trans