Abstract: To address the challenges faced by the Gold algorithm in reconstructing the low-energy resolution γ spectrum, such as the high number of iterations and the low counting rate for weak peaks, an iterative algorithm framework based on the Gold reconstruction algorithm was presented and it was extended to three novel iterative algorithms. The self-developed NaI(Tl) detector was selected as the experimental object. Firstly, the data required for the calculation of the detector’s non-negative response matrix were measured, and the corresponding non-negative response matrix was obtained. Secondly, the γ spectra of different radioactive sources were measured, and the convergence rate and recognition ability of the γ spectra were tested by using the detected γ spectra and the calculated non-negative response matrix, respectively. Three new iterative algorithms and Gold algorithm were compared and observed in the experiment of convergence rate. The convergence rate of each algorithm changes with the increase of the number of reconstruction iterations in the process of γ spectrum reconstruction. The reconstruction capabilities of these new iterative algorithms for the γ spectrum were analyzed and investigated through experiments. Based on a self-developed NaI(Tl) detector, low-energy resolution γ spectra were obtained through Monte Carlo simulations and actual measurements. The three proposed novel algorithms, in combination with a non-negative response matrix, were used to perform reconstruction analysis on the γ spectrum, and the results were compared with those obtained using the traditional Gold algorithm. The experimental show that the Mi-Gold (ratio factor m=1.9) algorithm significantly outperforms the Gold algorithm in terms of both convergence rate and reconstruction accuracy. The reconstructed spectral characteristic peaks are sharper and more distinct, with a relative peak position error of less than 1%, effectively enhancing the reconstruction effect of characteristic peaks in the low-energy region.