Deep learning (DL) tends to be widely applied in high-throughput imaging analysis. Its comprehensively application on image classification and prediction. However, general frameworks based on python package of TensorFlow have been few reported. It suggests more adjustments on frameworks are needed. This paper proposed, validated and integrated a framework of Google’s open-source Tensorflow computational library for high-throughput imaging. The software is much more friendly used and convenient. Its application on projects like leaf counting, age regression, and plant classification.