An image forming state detection device comprises a micro-lens array that is disposed at a position set apart from a predetermined focal plane of an image forming optical system by a specific distance and includes a plurality of micro-lenses arrayed with a predetermined pitch, a light-receiving portion array that includes a plurality of light-receiving portions each corresponding to one of the micro-lenses in the micro-lens array and receives an image on the predetermined focal plane via the individual micro-lenses, a signal string extracting means that extracts a pair of signal strings corresponding to images, formed with light fluxes having passed through different pupil areas of the image forming optical system, based upon light reception outputs obtained from the plurality of light-receiving portions, and an image forming state calculating means that calculates an image forming state at the image forming optical system by detecting an offset with regard to the phases of the pair of signal strings extracted by the signal string extracting means.