The remarkable sensitivity of ultrasonic waves to detect the early stages of material degradation stems from their inherent nonlinear nature. However, distinguishing the nonlinearity induced by the measurement system from that of the weak materials remains a persistent challenge in second harmonic generation methods. Ultrasonic wave mixing has emerged as a feasible solution for isolating nonlinearities originating from the measurement system. While bulk waves, surface waves, and guided waves have undergone extensive study in wave mixing, the potential of feature-guided wave (FGW) mixing remains largely unexplored. This paper investigates the mutual interaction of FGWs to generate second-order mixed harmonic waves, focusing specifically on weld joints. Acoustic energy trapped along structural features such as weld joints exhibits reduced decay compared to uniform plates, making them ideal for nonlinear characterization. We investigate the effectiveness of FGW mixing for characterizing material nonlinearities in welded joints, utilizing phase velocity matching and resonance criteria to enhance the generation of mixed harmonic components. Through three-dimensional numerical simulations and experimental tests, we showcase the generation of mixed harmonics induced by the mixing of FGWs. Our findings underscore the potential of FGW mixing as a novel approach for enhancing nonlinear-based damage detection techniques in welded joints.