The invention discloses an air cooling condenser ash accumulation state monitoring and cleaning control system based on a convolutional neural network and image recognition. The system is characterized by comprising a turbine and a steam exhaust allocation tube, wherein the turbine is communicated with the steam exhaust allocation tube; the steam exhaust allocation tube is communicated with a condensed water tank through a base tube of an air cooling condenser single row tube; the condensed water tank is communicated with a condensed water pump; a first visible light camera and a second visible light camera are arranged on the inner side of the base tube of the air cooling condenser single row tube and are positioned at the upper part of an air cooling fan; the first visible light camera and the second visible light camera are connected with an image acquirer; the image acquirer is connected with a monitoring host computer; the monitoring host computer is connected with a digital-modusconverter; the digital-modus is connected with a cleaning device control box; the cleaning device control box is connected with a compressed air cleaning device; the compressed air cleaning device issuspended on the lower side of a cleaning device walking truss; the cleaning device walking truss is arranged on the outer side of the base tube of the air cooling condenser single row tube and is connected with the steam exhaust allocation tube and the condensed water tank respectively. The invention further provides a scientific and reasonable regulation and control method.