Soft robots, with their excellent human-computer interaction features, offer a new tool for intestinal diseases detection. However, the existing soft robots rely on external control systems connected via cables, causing discomfort to patients during the detection process. Therefore, an earthworm-like cordless soft robot was proposed. The robot was equipped with a flexible pressure sensor and a wireless micro-camera, could perform controllable movements in intestinal-like environments, monitor changes of environmental pressure, and detect the foreign objects in the environment through visual. The robot was composed of a front-end radial motion actuator, a central axial motion actuator and a rear-end radial motion actuator, integrated with a compact control circuit board and 4 micro-pumps inside. By controlling the on/off of 4 air pumps via Bluetooth, the expansion and contraction of 3 motion actuators could be achieved, causing the robot to deform itself and enabling it to move in a intestinal-like environment. Experimental results demonstrated the robot's ability to move in both rigid and flexible tubes at a speed of 0.75 mm/min, and to effectively measure the changes of environmental pressure in real time. Additionally, the Fast RCNN (faster region convolutional neural net) algorithm was employed to process images captured by the robot, achieving precise identification of foreign objects in intestinal-like environments. This cordless soft robot has excellent human-computer interaction capabilities, providing a novel and patient-friendly approach for intestinal diseases detection.