One major drawback of using Translation Memories (TMs) in phrase-based Machine Translation (MT) is that only continuous phrases are considered. In contrast, syntax-based MT allows phrasal discontinuity by learning translation rules containing non-terminals. In this paper, we combine a TM with syntax-based MT via sparse features. These features are extracted during decoding based on translation rules and their corresponding patterns in the TM. We have tested this approach by carrying out experiments on real English–Spanish industrial data. Our results show that these TM features significantly improve syntax-based MT. Our final system yields improvements of up to +3.1 BLEU, +1.6 METEOR, and -2.6 TER when compared with a stateof-the-art phrase-based MT system.