An image display device of the present disclosure includes an image light generating device, a first, a second, a third, and a fourth optical unit. A first intermediate image is formed between the first and the third optical unit. A pupil is formed between the second and the fourth optical unit. A second intermediate image is formed between the third and the fourth optical unit. An exit pupil is formed at an opposite side of the fourth optical unit from the third optical unit. The image light generating device includes a first, a second, a third light emitting panel, and a color synthesis element. The color synthesis element is constituted of a cross dichroic prism including a first and a second dichroic film that intersect with each other. Each of the first and the second dichroic film does not have a polarization separation characteristic.