The invention discloses a hot spot fault diagnosis method of a photovoltaic cell panel, and belongs to the field of fault detection of a photovoltaic power generation system. The hot spot fault diagnosis method comprises the steps of firstly, regularly acquiring detection data of each branch in a photovoltaic array, and building an isolation forest-based hot spot fault estimation model of the photovoltaic array, wherein little and different data can be analyzed from a large amount of data by the model so that a fault branch is determined and hot spot fault score is given; and finally, positioning a cell panel with a hot spot fault in the branches by employing a cross measurement method.