Non-pharmacological therapies, especially the physical maneuvers, are viewed as important and promising strategies for reducing syncope recurrences in vasovagal syncope (VVS) patients. We observed the efficacy of a modified Valsalva maneuver (MVM) in VVS patients. 72 VVS patients with syncope history and positive head-up tilt table testing (HUTT) results were randomly divided into conventional treatment group (NVM group, n = 36) and conventional treatment plus standard MVM for 30 days group (MVM group, n = 36). Incidence of recurrent syncope after 12 months (6.5% vs. 41.2%, P<0.01) and rate of positive HUTT after 30 days (9.7% vs.79.4%, P<0.01) were significantly lower in MVM group than in NVM group. HRV results showed that low frequency (LF), LF/ high frequency (HF), standard deviation of NN intervals (SDNN) and standard deviation of all 5-min average NN intervals (SDANN) values were significantly lower in the NVM and MVM groups than in the control group at baseline. After 30 days treatment, LF, LF/HF, SDNN, SDANN values were significantly higher compared to baseline in MVM group. Results of Cox proportional hazard model showed that higher SDNN and SDANN values at 30 days after intervention were protective factors, while positive HUTT at 30 days after intervention was risk factor for recurrent syncope. Our results indicate that 30 days MVM intervention could effectively reduce the incidence of recurrent syncope up to 12 months in VVS patients, possibly through improving sympathetic function of VVS patients.