Determining the harmonic contribution of each harmonic source is conducive to improving the harmonic management of the power system. The main problem with harmonic contribution calculation is determining the equivalent system harmonic impedance. The previous harmonic impedance calculation method can only calculate the impedance value when the harmonic fluctuation is significant, and the calculation fails when the harmonic fluctuation is small. Therefore, a new method called difference recurrence estimation is proposed to construct the objective function from impedance parameters. The advantages of the proposed method are low data requirement, high estimation accuracy, and excellent tracking performance. An improved genetic algorithm (GA) is presented by integrating an adaptive crossover operator and population intervention into the standard GA to improve the global searching impedance parameter ability. Based on these parameters, harmonic impedance and harmonic contribution are calculated accurately. The superiority of the proposed model is verified through simulation and field cases.