Advances in wearable technology provide opportunities to collect biomechanical data in real time and non-lab settings. However, there are currently no standards for best practice use of wearable sensors for gait applications. PURPOSE: Provide best practice recommendations for use of wearable accelerometers (WA) in gait biomechanics. METHODS: Literature was reviewed to determine appropriate WA range for gait biomechanics and establish procedures for calibrating and processing WA data for gait applications. Drop tests of three, simultaneously initialized, commercially available WAs were performed to determine signal time synchronization and 1 g acceleration accuracy. WAs were secured to different lower limb locations (pelvis, knee, ankle) and walking and running trials performed. Peak acceleration magnitude and timing were compared within and among WAs by location and gait type. Vertical ground reaction force (GRFvert) was estimated using a regression model, developed based on pelvis acceleration data, to determine the sensitivity of the GRFvert estimates to WA placement. RESULTS: Peak lower limb accelerations can exceed 25 g during running. WAs initialized at the same time had significant temporal differences (up to 1.06 s). Accelerations during freefall were within 17 % of 1 g. After synchronizing WA signals based on the drop test results, there were no significant differences in WA magnitude and timing among three WAs located around the right iliac crest, but there were significant acceleration differences among WAs located at the right iliac crest, knee, and ankle. Walking and running peak GRFvert estimates based on accelerations of the iliac crest differed from estimates based on accelerations of the knee (-5 ± 28 N and -142 ± 80 N, respectively) and ankle (261 ± 28 N and -274 ± 66 N, respectively). CONCLUSIONS: WAs should be: (1) selected to measure a range greater than 25 g, (2) calibrated to ensure accuracy, (3) manually time synched if using multiple sensors, (4) placed carefully though exact placement is not critical to anatomical site acceleration estimates near the hip, and (5) located at the same anatomical site that was used to develop an acceleration-dependent model if the goal is to use that model to estimate a particular quantity (i.e. GRFvert).