Abstract: The accurate indication of control rod positions plays a pivotal role for remote real-time monitoring and self-intelligent operation of nuclear energy systems. In current pressurized water reactor systems, control rods operate in harsh environments under high temperature, high pressure, and high radiation. Typically, non-contact rod position detectors or position encoders are used to monitor the position of control rods. Nevertheless, numerical methods for predicting control rod positions have gradually emerged, showcasing notable advantages. The prismatic super-safe gas-cooled reactor (referred to as gas-cooled micro reactor) is one of the movable advanced micro nuclear reactor types, employing TRISO dispersed fuel with excellent safety features, cooled by helium gas, and possessing inherent safety characteristics. Based on the structural characteristics of the gas-cooled micro reactor, a harmonic parameterization method was proposed. Firstly, the process began with the preparation of high-order harmonic libraries at different rod positions (30-170 cm) using the Monte Carlo method. Secondly, the most influential high-order harmonics, identified through power reconstruction effectiveness, were selected. Thirdly, harmonic parameterization fitting calculations were performed after completing the harmonic selection. The main content of harmonic parameterization was to obtain continuous high-order harmonics with rod position by fitting the high-order harmonics of discrete rod positions using a polynomial fitting method. Fourthly, after establishing the high-order harmonics as a function of control rod positions, the study statistically derived the response matrix of external detectors using the Monte Carlo method. Finally, integrating this response matrix with the least squares method facilitates the resolution of the detector response equation. This methodology culminates in the successful prediction of control rod positions via the minimum residual approach. Simultaneously, corresponding high-order harmonics and core power reconstruction could be obtained. Validation through the gas-cooled micro reactor model demonstrates the efficacy of the harmonic parameterization method in accurately predicting control rod positions across 15 sets (30-170 cm) with an average deviation of 1.6 cm. The maximum deviation occurs at a control rod position of 70 cm, with a deviation of 3.2 cm, indicating a high prediction accuracy. Compared to fixed high-order harmonics, the maximum root mean square relative error of nodal power reconstruction decreases from 6.1% to within 3%, and the maximum relative error decreases from 14.5% to within 5% using the harmonics obtained from the harmonic parameterization method. These findings validate the feasibility of using harmonic parameterization to predict control rod positions and enhance the accuracy of core power reconstruction. This approach provides a new avenue for future online monitoring and fault diagnosis of gas-cooled micro reactors.