The genetic code recognition system in modern ribosome is a very complicated system and it is not easy to touch the core of the problem. We built up a simplified chemical model which only consists of phosphorous compounds, silylated amino acids and silylated nucleosides to study the origin of the genetic code. It was found that the yields of the “translation products”, namely dipeptides, for each of the six amino acids (Phe, Trp, Tyr, His, Val and Leu) behaved differently with the presence of different nucleosides. The proposed mechanism involves a five‐membered cyclic pentacoordinated nucleoside‐phosphoric acid‐amino acid anhydride (NPA) as a key intermediate. This reaction mechanism could explain why nucleosides may have affected the synthesis of dipeptides in processes that eventually led to translation and the genetic code.