The industrial scene images, affected by various factors, always have complex background, so a single feature cannot achieve satisfactory result for classification. In this paper, a method using the image color and texture features through the Bp_adaboost method to classify the industrial instrument image is proposed. It first extracts the low order color moment features in HSV space, and then extracts the texture features based on the gray level co occurrence matrix. Finally, the 17 dimensional integrated feature vector for industrial instrumentation image is adapted to the Bp_adaboost classification learning and testing. Experimental results show that the method can achieve better classification results for the industrial level control system instrumentation equipments and level container images.