Background: The optimal treatment sequence in current standard care for patients (pts) with anti-EGFR antibodies-naïve metastatic colorectal cancer (mCRC) is cetuximab (C) followed by regorafenib (R). The objective of this randomized phase II trial is to evaluate efficacy and safety of the therapeutic sequence of R followed by C compared with that of C followed by R for mCRC pts. Methods: Pts with KRAS exon 2 wild-type mCRC after failure of fluoropyrimidine, oxaliplatin, and irinotecan were randomized to receive sequential treatment with R followed by C ± irinotecan (R-C arm) or reverse sequence (C ± irinotecan followed by R; C-R arm). The primary endpoint was OS. Key secondary endpoints included PFS with initial treatment (PFS1), PFS with second treatment (PFS2), safety, and QOL. The exploratory endpoint was serial biomarker analyses including oncogenic alterations from ctDNA or multiple serum proteins. Results: Between November 2013 and September 2016, 101 pts (51 in the R-C arm and 50 in the C-R arm) were randomized and eligible for efficacy analysis. Baseline characteristics were well-balanced in both arms. Bevacizumab had been previously administered in 96% and 98% pts in R-C and C-R, respectively. Sequential treatment was succeeded in 86% pts in both arms. After a median follow-up of 29.0 months with 81 death events, median OS in R-C and C-R were 17.4 and 11.6 months, respectively (stratified logrank P = 0.0293), with a hazard ratio (HR) of 0.61 for OS (95% CI, 0.39–0.96). HR for PFS1 (R in R-C versus C in C-R) was 0.97 (95% CI, 0.61–1.54), whereas that for PFS2 (C in R-C versus R in C-R) was 0.29 (95% CI, 0.17–0.50). No unexpected safety signals were observed in both arms. OS results in RAS/RAF wild type populations with ctDNA analysis at study entry (n = 86) were similar to those of overall population. Conclusions: This was the first randomized study to compare the two therapeutic sequences of R and C for mCRC, suggesting R followed by C is the preferred sequence. Comparable PFS1 and remarkable PFS2 improvement may lead to a longer OS in R-C sequence. The results of detailed biomarker analysis would be also reported. Clinical trial information: UMIN000011294.