Children with cerebral palsy (CP) exhibit head instability during simple overground walking, which may comprise sensory input and reduce stepping accuracy. Investigations of head stability during more challenging tasks, where fall risk may be increased, are limited. This study explored differences in head stability between ambulatory children with hemiplegic CP (N = 9) and diplegia (N = 9) (GMFCS I and II) and typically developing (TD) children (N = 8) during a targeted stepping task. All children completed five trials stepping into two successive rectangular floor-based targets whilst walking along an 8 m walkway. Three-dimensional motion capture enabled calculation of head stability and foot placement within and before each target. A two-way mixed-design ANOVA compared differences between all groups and target approach. Children with diplegic CP showed greater sagittal, frontal, and resultant head-to-laboratory and head-to-trunk head instability compared to children with hemiplegic CP and TD children. Anteroposterior foot placement error was significantly greater in children with hemiplegic CP (8.5 ± 5.0 cm) compared to TD children (3.8 ± 1.5 cm). Group differences in head instability were not consistent with group differences in foot placement error. To better understand how head instability might affect fall risk in children with CP, more challenging environments should be tested in future.