A method to investigate optical path turbulence based on laser spot distortion and a convolutional neural network (CNN) is proposed. Utilizing the CNN, we evaluated the spot distortion of laser beams resulting from airflow disturbance in space propagation. As a result, details of turbulence on the beam propagation path can be obtained. Experimental results demonstrate a high correlation between the evaluation parameter and the turbulent intensity (wind speed) measured by an anemoscope. The proposed method provides a turbulence analysis with short distance, high speed, and low cost.