Reactive radicals can significantly affect the timely exposure of the catalyst’s active site and lead to a significant impact on reaction rates, especially for long-lived radicals. Here, two types of metal-free polymers are designed according to the theoretical calculations: one without acetylene, which produces more long-lived superoxide radicals, and another with acetylene, which produces more short-lived hydroxyl radicals. Subsequent experimental characterization revealed that the metal-free polymer containing acetylene (P1-E) outperformed the acetylene-free polymer (P1) in terms of photon absorption, the separation of photogenerated charges, and the photoreduction of heavy metals (Cr(VI)). However, when tetracycline was introduced as a radical scavenger, the continuous consumption of long-lived superoxide radicals resulted in a 43-fold increase in the rate constant (k) of Cr(VI) photoreduction in the P1 system (from 0.00309 to 0.13248 min). In contrast, in the P1-E system dominated by short-lived hydroxyl radicals, the rate of Cr(VI) photoreduction increased by a factor of only 3 (from 0.00623 to 0.02111 min). This significant transformation can be attributed to the prolonged occupation of active sites on the catalyst surface by long-lived free radicals, thereby inhibiting the efficiency of the reduction reaction. However, for short-lived free radicals, their rapid annihilation exposes more active sites on the catalyst. Therefore, in the coexisting system of tetracycline and Cr(VI), there are distinct differences in the performance of P1 and P1-E catalysts for the photoreduction of Cr(VI). The same phenomenon was also observed with the addition of other contaminants (p-nitrophenol, levofloxacin, and carbamazepine). This study highlights that the removal of long-lived radicals can significantly improve the photocatalytic performance and offers new insights for enhancing the efficiency of photocatalytic reactions.