Machine translation is one of the oldest and hardest problems in artificial intelligence. It is studied as a subfield of natural language processing or computational linguistics. Although machine translation has a long history, full automatic translation with high quality seems hard to achieve at least in the near future. Nowadays due to the increase in computing power and the accessibility of huge date in the Internet, the field has taken a new direction and started to grow rabidly. Translation is a process of transferring the meaning of words or text to another language which involves decoding the meaning of the source language and then re-encoding that into the target language. Many systems have been proposed in order to tackle this problem such as rule-based, example-based, statistics-based and other systems. Evaluating these systems is another difficult issue since there is no one correct translation of a sentence and it is subjective to humans' judgement. Machine translation evaluation is currently very active research and has many debatable topics.