A fluorescence observation apparatus, including: a light source configured to irradiate a subject with reference light and excitation light; and one or more processors including hardware, wherein the one or more processers are configured to implement: a fluorescence-image generating portion configured to generate a fluorescence image; a reference-image generating portion configured to generate a color reference image; an extraction portion configured to extract a fluorescence region from the fluorescence image; a motion-effect generating portion configured to generate a frame image bringing about an effect that is visually time-varying with a variation quantity depending on a gradation value of the fluorescence region, in a region corresponding to the fluorescence region; and a synthesis portion configured to add the frame image to any of a plurality of color-component images constituting the reference image to synthesize a synthetic image.