Conventional structural health monitoring (SHM) using ultrasonic guided waves (UGW) typically depends on baseline signals. However, baseline signals are prone to environmental variations and sensor instability, which compromises the accuracy and reliability of SHM. To address this challenge, this study proposes a novel baseline-free damage imaging method based on similar paths. Using a circular sparse array sensor, the method calculates the root mean square error (RMSE) of similar path signal matrices to adaptively identify pseudo-baseline signals. These signals are then used to compute six damage indices for subsequent damage imaging. Numerical simulations of single and dual damages in aluminum plates were conducted to validate the method. Results demonstrate that the proposed similar-path-based pseudo-baseline signals can effectively replace traditional baseline signals, enabling precise identification and location of 10 mm diameter single and dual damages. Furthermore, the method maintains high location accuracy across different β-parameter values of the Reconstruction Algorithm for Probabilistic Inspection of Damage (RAPID), thereby offering a more robust and reliable solution for SHM.