Objective To investigate the clinical effect of paraquat (PQ) detoxification recipe combined with continuous hemoperfusion (HP) in the treatment of patients with acute paraquat poisoning (APP) and clinical significance of soluble CD14 subtype (sCD14-st, Presepsin). Methods A prospective randomized controlled trial was conducted. 152 patients with moderate APP admitted to Department of Emergency Medicine of Harrison International Peace Hospital Affiliated to Hebei Medical University from July 2013 to June 2017 were enrolled, and they were randomly divided into three groups. The patients in HP group (group A, n = 35) only received 2-hour HP for 3 times, 8 hours each time, those in PQ detoxification recipe combined with HP group (group B, n = 50) received PQ detoxification recipe 1 (once per 2 hours until no PQ component was found in faeces) and 2 (3 times a day for 14 days) beside HP. The others in PQ detoxification recipe combined with persistent HP group (group C, n = 67) received continuous HP until the PQ component in serum was not detected. The parameters of organ function and inflammatory factor, and blood Presepsin and PQ contents were determined before and after treatment. The curative effect and 28-day mortality were recorded. The correlations between serum Presepsin level and PQ content as well as 28-day mortality were analyzed with Pearson correlation analysis. Receiver operating characteristic curve (ROC) was plotted to analyze the predictive value of Presepsin on prognosis. Results The total effective rate of group C was significantly higher than that of groups A and B [70.1% (47/67) vs. 34.3% (12/35), 54.0% (27/50)], and 28-day mortality was significantly lowered [29.8% (20/67) vs. 65.7% (23/35), 46.0% (23/50), both P < 0.05]. There was no significant difference in alanine aminotransferase (ALT), MB isoenzyme of creatine kinase (CK-MB), serum creatinine (SCr), C-reactive protein (CRP), tumor necrosis factor-α (TNF-α), and interleukins (IL-6 and IL-10) before treatment among the three groups. Five days after treatment, the above parameters in the three groups were increased as compared with those before treatment, but the increase degree in group C was the lowest. At 7 days after treatment, the parameters were decreased, especially in group C. There was no significant difference in serum Presepsin and PQ levels before treatment among the three groups. With the prolongation of treatment time, the Prespsin levels in groups A, B, and C were increased, and peaked at 12 hours (μg/L: 4.28±0.20, 3.87±0.25, 3.53±0.23), then gradually decreased,and the PQ contents were lower than those before treatment from 8 hours (mg/L: 1.76±0.12 vs. 2.12±0.17, 1.57±0.08 vs. 2.24±0.16, 1.25±0.10 vs. 2.14±0.18), with a time dependence pattern, especially in group C (all P < 0.05) . Correlation analysis showed that blood Presepsin level was positively correlated with PQ content and 28-day mortality (r1= 0.917, r2= 0.864, both P = 0.001), suggesting that the higher the PQ content was, the higher the Presepsin level, and the higher the 28-day mortality was. ROC curve analysis showed that the area under ROC curve (AUC) of Presepsin predicting 28-day mortality was 0.863; when the cut-off value was 1.22 μg/L, the sensitivity was 83.3%, the specificity was 81.4%, the positive predictive value was 77.46%, and the negative predictive value was 86.42%. Conclusions Early administration of PQ detoxification recipe combined with continuous HP treatment can effectively reduce Presepsin level, decrease the mortality of patients with moderate APP, improve the prognosis. Presepsin can assess the prognosis of patients with APP.