The poem code is a simple and insecure cryptographic method.

The method works by the sender and receiver pre-arranging a poem to use. The sender chooses a set number of words at random from the poem and gives each letter in the chosen words a number. The numbers are then be used to conceal the clear text of the message using double transposition. To indicate to the receiver which words had been chosen an indicator group is sent at the start of the message.

The essential weakness of the code is that if one message is broken by cyptoanalytical attack then future messages will be vulnerable if the source poem is identified. Since the poem used has to be memorable they can often be taken from classic sources (the works of Racine, Moliere, Keats or similar), making identifying the source extremely easy.

The weakness of the method became clear during WW II and the operations of the SOE. At the insistence of their codes officer, Leo Marks, the SOE began using original compositions to give some added protection, but also adopted other more secure methods such as the one-time pad.