Background: Polysomnography (PSG) is the gold standard for sleep bruxism (SB) assessment. High economical costs, complex technical equipment, and unfamiliar laboratory setting limit its use in children. Aim: To determine the night-to-night variability of electromyography (EMG) episodes during a five-night recording with the GrindCare Measure (GCM), and the agreement in the assessment of masticatory muscle activity (MMA) between GCM and PSG in children. Design: Forty-seven children from clinics of Universidad CES participated. Each participant was assessed with GCM for five consecutive nights. The last night, children underwent a single-night PSG study, together with the GCM. Spearman correlation coefficients were used to analyze data. Results: The frequency of SB occurrence was ‘sometimes’ in 12 (25.5%) and ‘usually’ in 19 (40.4%) children. Simultaneous measurements with GCM and PSG obtained during the fifth night of measurement were not significantly correlated. Correlation between GCM total EMG episodes and EMG episodes/h and PSG total SB episodes, SB episodes/h, total bursts and burst/h measured with PSG was also not significant. Conclusion: EMG measurement with GCM was not accurate to detect PSG/SB in children. There was not advantage of multiple assessment for five nights with GCM, reducing the impact of night-to-night EMG episodes’ variability on the GCM/PSG correlation.