The objective of this project is to show through a statistical model of a completely randomized design in a crop of grass and forage (cut grass) which was carried out in vanillo, this is located at km 48 via the triumph of the province of Guayas. The variety that we had was female elephant grass (Pennisetum purpureum Schum.) and purple elephant grass (Pennisetum purpureum Schumacher) chosen by Mr. Carlos Ramirez, who directs the Pasture and Forage course which is related to the experimental design course directed by Ms. Segress Garcia. Thanks to an analysis of variance design model it will be possible to obtain the exact results. In order to carry out a process we must make some calculations, which are: the sum of square of the treatments, sum of square of the blocks and the sum of total square, with respect to the calculated F must always be greater than the tabulated F which gives us if there is significant difference or not but with this we still cannot determine which is the best treatment without first doing a test of means by Fisher, otherwise the completely randomized design would only end up there.