The seven-wire strands are the crucial components of prestressed structures, though their performance inevitably degrades with the passage of time. The ultrasonic guided wave methods have been intensely studied, owing to its tremendous potential for full-scale applications, among the existing nondestructive testing methods, for evaluating the stress status of strands. We have employed the theoretical and finite element methods to solve the dispersion curve of single wire and steel strands under various boundary conditions. Thereafter, the singular value decomposition was adopted to work with the simulated and experimental signals for extracting a feature vector that carries valuable stress status information. The effectiveness of the vector was verified by analyzing the relationship between the vector and the stress level. The vector was also used as an input to establish a support vector regression model. The accuracy of the model has been discussed for different sample sizes. The results show that the fundamental mode dispersion curve offset on the high-frequency part and cut-off frequency increases as the boundary constraints enhance. Simulated and experimental results have demonstrated the effectiveness and potential of the proposed support vector regression method for evaluating the stress level in the strands. This method performs well even at low stress levels and the reliability can be enhanced by adding more samples.