BACKGROUND: When treating knee osteoarthritis (KOA), rehabilitation exercise is often combined with low-level (LLLT) or high-intensity laser therapy (HILT). However, the effectiveness of these combinations is still uncertain. OBJECTIVE: To examine the effects of LLLT or HILT combined with rehabilitation exercise (LLLT+E or HILT+E) on pain, stiffness and function in KOA. DATA SOURCES: Five databases (PubMed, Web of Science, CINAHL, PEDro and SPORTDiscus) were searched for relevant studies published up to 31 December 2019. STUDY SELECTION: KOA randomised controlled trials, LLLT+E or HILT+E against exercise alone with or without placebo laser, reported clinical outcomes, human studies and English language. DATA SYNTHESIS: Methodological quality was assessed by the PEDro and Cochrane risk-of-bias scales, and the meta-analysis was undertaken using RevMan 5.3. RESULTS: Of the 10 retrieved studies, six investigated LLLT+E, three on HILT+E, and one evaluated both. All the studies had high PEDro scores. However, as most of the studies employed a single type of laser therapy, only indirect comparison of LLLT+E and HILT+E was possible. This study found all treatment modalities were effective in reducing KOA symptoms. Interestingly, relative to control, the meta-analysis showed significant improvements in knee pain, stiffness and function for the HILT+E. CONCLUSION: Both LLLT and HILT are beneficial as adjuncts to rehabilitation exercise in the management of KOA. Based on an indirect comparison, the HILT+E seems to have higher efficacy in reducing knee pain and stiffness, and in increasing function. To confirm this finding, a direct comparative investigation of the two types of laser therapy may be necessary. Systematic review registration PROSPERO ID CRD42020206228.