To safely walk in a community environment requires dual cognitive–walking ambulation ability for people with Parkinson’s disease (PD). A past study showed inconsistent results on cognitive–walking performance for PD patients, possibly due to the various cognitive tasks used and task priority assignment. This study designed cognitive–walking tests that used executive-related cognitive tasks to evaluate patients with early-stage Parkinson’s disease who did not have obvious cognitive deficits. The effect of assigning task prioritization was also evaluated. Sixteen individuals with PD (PD group) and 16 individuals without PD (control group) underwent single cognitive tests, single walking tests, dual walking tests, and prioritizing task tests. Three types of cognitive, spatial memory, Stroops, and calculation tasks were employed. The cognitive performance was evaluated by response time, accuracy, and speed–accuracy trade off composite score. The walking performance was evaluated by the temporal spatial gait characteristics and variation in gait. The results showed that the walking performance of the PD group was significantly worse than the control group in both single and dual walking conditions. The group difference in cognitive performance was shown in composite score under the dual calculation walking task but not under the single task. While assigning priority to walking, no group difference in walking was observed but the response accuracy rate of PD groups declined. This study concluded that the dual task walking test could sharpen the cognitive deficits for early-stage PD patients. The task priority assignment might not be recommended while testing gait deficits since it decreased the ability to discriminate group differences.