In this paper, we design a computer vision frame-works for fish size measurement. By using computer vision approach, in the measurement process becomes easier, without touching the object and faster. So that, the freshness, sterile and hygiene of the fish will be maintained. Fish size measurement is usually done manually, and this takes a very long time and is not accurate. We overcome this problem by designing a framework based on computer vision, where the camera will record real-time video, then the video will be processed through a single board computer. The single board computer have a several processes, starting from uploading the ArUco detector library to drawing boundaries on objects. Then the measurement results will be issued through the printer and RS232 rejector. Where the RS232 rejector is connected to a PC/laptop, so the measurement results can be displayed on the PC/laptop screen. With the size of the ArUco marker of 8x8 cm, the circumference is 32cm. Where the circumference of the ArUco marker functions as a reference value that is used to perform error testing of the proposed system. Experiment results shows the system have a good performance with average error is 0 % in width and 0.06 % in length.