Path prediction from an incomplete trajectory path is an important problem of practical relevance and finds applications in understanding and planning navigation, predicting missing trajectory, etc., of ships in the sea. Historical data of trajectories can be of significance in hypothesising the trajectory by recommending from the learnt trajectory pattern. Towards this, the paper brings out a novel approach ‘GETSSL’ to predict missing segment of a trajectory. GETSSL refers to ‘Geohash Encoding of Trajectory and Sequence-to-Sequence Learning’. In this approach, we transform ship trajectory into sequence of words by encoding the trajectory using Geohashes. The encoded trajectory is used to train Encoder-Decoder Deep Learning Model under the sequence-to-sequence framework. We learn embeddings on geocoded trajectories and use during sequence prediction task. Towards this work experiments have been carried out using open source AIS dataset and the applicability of the formulation and results have been validated and found to be promising.