To apply the LaBr3(Ce) detector with a higher full-energy peak detection efficiency and larger activity detection range to the measurement of low- and intermediate-level radioactive waste (LILW), we examined the problem of poor energy resolution. We analyzed the characteristics of key nuclide energy spectra and established an adaptive filter function to form a deconvolution energy spectrum analysis method for intelligent nuclide identification and quantitative calculation. Compared with the traditional deconvolution method, the average overlapping peak resolution was improved by 74.5%, achieving the qualitative analysis and quantitative calculation of all nuclides. By using the novel method to analyze the experimental spectrum of Co-60 and Cs-137, the results showed that the nuclide was accurately identified with a mean relative error of 0.85% for the peak area calculations. The analysis of the spectra of the compositions of the five LILW nuclides was obtained from Monte Carlo simulations using different methods. The results showed that, for the overlapping peaks with peak area ratios of 1:1:1:1:1, compared with the traditional deconvolution method, the peak area calculation error of this method was reduced from 73.09% to 2.12%. For the overlapping peaks with the peak area ratios of 5:1:2:3:1 and 5:1:20:10:1, the traditional spectrum-unfolding method could not resolve the weak peaks, and the maximum energy deviation was 8 keV. This method achieved accurate identification of all energies, and the error of the weak peak was 7.19%, and the error of the other peaks was less than 2.23%. The results of this study provided key technical support for the application of LaBr3(Ce) detectors to the nuclide measurement of LILW.