The accurate location of radiated passive intermodulation (R-PIM) source activated by antennas has become a tough problem in modern multicarrier wireless communication systems. In order to precisely locate multiple harmful R-PIM sources, we develop a novel imaging method called the cascaded time-reversal multiple signals classification (CTR-MUSIC). This method multiplies a group of time-reversal multiple signals classification (TR-MUSIC) pseudo-spectra together to form the new pseudo-spectrum. The cascaded multiplication highly brightens the pseudo-spectrum peaks to easily identify the positions of the R-PIM sources and suppresses the remained pseudo-spectrum approximately to 0. By combining the multiple-static data matrices (MDMs) optimized by the weak-correlation filter (WCF), the proposed method exhibits a good ability in removing the redundant frequency information. The imaging computational complexity, positioning error, and positioning resolution of CTR-MUSIC are analyzed by numerical simulations. Compared with the existing ultra-wideband TR-MUSIC (UWB-TR-MUSIC), the results show that the proposed method improves the positioning accuracy, reduces computation consumption several times, and decreases the number of detecting antennas required for the same positioning accuracy. Experimentally, the R-PIM source is located with the positioning error of 0.04 wavelength at the signal central frequency of 1.7 GHz that proves the feasibility of positioning R-PIM source using the proposed method.