Abstract: In view of the status of low visual inspection accuracy for pumping unit indicator diagram, complex training in automatic identification neutral network and low computation speed, this studied the automatic identification technique for indicator diagram based on fractal box dimensions, verified the viability of indicator diagram by fractal box dimensions theory, provided automatic identification steps and identified the indicator diagram under four working conditions: normal and insufficient liquid supply by rodless pumping unit, leaking at traveling valve and tubing leaking. The research result shows that the indicator identification technique based on fractal box dimensions can accurately identify the indicator diagrams under the above four working conditions and more accurately when at low noise level; when the noise level is relatively high, noise reduction processing needs to be performed in order to ensure accurate automatic identification. The research presents a new method for automatic identification of pumping unit indicator diagram and is of great significance to improve the level of automation of pumping unit running monitoring and intelligent analysis.