The mathematics of belief functions can be handled with the use of the matrix notation. This representation helps greatly the user thanks to its notational simplicity and its efficiency for proving theorems. We show how to use them for several problems related to belief functions and the transferable belief model.