In the Philippines, two of the most spoken dialects are Cebuano and Hiligaynon. Moreover, Cebuano and Hiligaynon are closely related in the sense that both languages share words but have different meaning according to context. The study aimed to create a translation model for Hiligaynon to Cebuano by applying recurrent artificial neural networks with long short-term memory. Two neural networks were developed, one for encoding source sentences and one for decoding target sentences in a manner following sequence-to-sequence learning. The highest accuracy achieved was when the model was trained at 150 epochs and 4000 sentences, yielding a BLEU score of 0.265579454. It can be concluded that a neural machine translation model can be created given sufficient training data.