Abstract: In recent years, "the Belt and Road" has become an important strategy to the development of China's foreign trade. However, owing to the cultural difference among the involved countries from Asia, Africa and Europe, which have many ethnic groups, the existing barriers still hinder bilateral trade. This paper analyzed the effect of language on the entire and multi-industry trade between different countries involved in "the Belt and Road" by using the gravity model and applying CEP II language database. The empirical results show that the direct communication skill (common spoken language) has the largest impact on the "The Belt and Road" bilateral trade. Other factors, according to its importance level to bilateral trade of "The Belt and Road", are ethnic affiliation (common native language), indirect communication that officially supported by government (common official language) and traditional indirect communication (linguistic proximity). Meanwhile, with the complexity of products increases, the empirical results show that the language effect on multi-industry trade becomes large. For homogeneous products, it relies on low-cost communication due to the low value-added, and therefore the importance of language is very low. For intermediate products, the importance of the language has increased. For heterogeneous products, common native language is largest impact factor on the trade of "the Belt and Road". Others followed by common spoken language, common official language and linguistic proximity, reflect the importance of linguistic factors on heterogeneous products trade of "the Belt and Road".