Squid industry is an important part of my country's marine economy, but the squid processing industry is limited by manual labor, which seriously restricts the upgrading of the industry. In order to realize the automation of the separation process of squid carcass, viscera, fins and heads, this study designed a carrier with a "fin-clamping and neck-hanging" fixed mode for squid according to the structural characteristics of squid, and designed a squid automatic feeding module, squid yellowing, yellowing, finning and head removal structure by analyzing the characteristics of manual labor. The automatic feeding module realizes squid posture recognition based on the Faster-RCNN neural network model, realizes squid grabbing point solution based on the fast DT algorithm, and uses collaborative arms to complete the automatic feeding of squid; the yellowing, yellowing, finning and head removal structure fusion control program realizes human-like operation. The test results in the squid processing plant show that the fully automatic removal equipment for squid by-products can realize the automatic feeding, visceral removal, finning and head removal of squids in the specification range of 100-400g, with a processing efficiency of 45 pieces/minute, and a single device can save 3 people. Research has shown that the fully automatic squid by-product removal equipment can simulate humans to complete the work of removing squid viscera, fins and heads, achieving the same operating effect, reducing labor intensity while improving work efficiency. Its promotion and application can improve the intelligence level of the squid processing industry.