A scanning type image display apparatus is disclosed which is capable of forming an exit pupil with a shape and a size which facilitate observation. The apparatus includes a scanning unit, a first optical system introducing a light flux to the scanning unit, a second optical system converging the light flux from the scanning unit, a diffractive optical element receiving the converged light flux, and an ocular optical system. Two directions in diametral directions of the light flux are first and second directions. The optical element increases the divergent angles of an emerging light flux therefrom in the first and second directions as compared with the convergent angles of an incident light flux thereinto in the two directions and increases the divergent angle of the emerging light flux in the first direction as compared with the divergent angle thereof in the second direction.