The development of noise reduction and silencing technology has brought great difficulties to underwater target detection, and more target characteristics need further studying. When a submerged target travels through density-stratified environment, the fluid will oscillate behind the target owing to gravity and buoyancy and generate internal waves, which are often referred to as source-generated internal waves. These internal waves are difficult to eliminate, which can cause the sound speed profiles to fluctuate. Therefore, these internal waves are expected to be effective for detecting underwater target. In this paper, the fluctuations of the received sound passing through the internal waves produced by a moving sphere are investigated. A typical shallow stratified environment is set up, and internal wave fields generated by a sphere moving in many horizontal directions are simulated. According to the simulation results, these internal wave fields have a much wider range than the scenario of the target body. Based on the relationship between the amplitude of the internal wave and the variation of sound speed, range–dependent sound speed profiles are constructed, and model based on ray acoustics is used to analyze the aberration strength of passing sound fields. Results show that the strength aberration is inversely proportional to the target passing angle, and these characteristics can be covered by the background. Focusing on this problem, an extraction method based on principal component analysis with sliding window is then proposed. The uncorrelation between the disturbance of internal wave and background signal is utilized, and interference is suppressed by removing the component in No.1 principal component space, and retaining the No.2–No. ksubspace. Detection can be executed based on multi period received data from single hydrophone. A lake experiment is conducted to verify the performance. A detection scenario of single source and single receiver is established, and the AUV target crosses source–receiver line multiple times. The research results show that the detection scheme based on the acoustic aberration of source-generated internal wave has potential for underwater target detection, possessing the advantages of wide coverage and high robustness. Data on multi depths are processed to show that the detection performance is dependent on the depth of system. Since the acoustic strength variations are derived form local disturbance in channel, the proposed method may be affected by severe environment fluctuation, and further research is still needed.