Machine translation has seen a paradigm shift with the introduction of Neural Networks. The improvement has enabled multiple pathways in establishing communication among natives of different language speakers. In this cue the techniques are ever-growing and efficient than the prior ones. In this chapter the use of bidirectional long short-term memory with attention mechanism has been utilized to improve the accurateness and fluency of English-to-Urdu translation. The study is based on different dataset with a total of 793,358 English–Urdu parallel phrases of different categories containing general and religious content. The suggested model has an enhanced BLEU score, indicating its increasing efficiency in enhancing machine translation system’s accuracy and fluency.