Aim:  To examine whether the Frontal Assessment Battery is associated with the immediate effects of physical therapy on gait disturbance in patients with Parkinson's disease. Methods:  A total of 18 patients with idiopathic Parkinson's disease (Hoehn and Yahr stage range 3–4) who were able to ambulate independently and who were not demented were included. Patients were divided into two groups on the basis of Frontal Assessment Battery scores: the high score group (score ≥13, n = 11) and the low score group (score ≤12, n = 7). A 3‐D motion analysis system was used to acquire gait parameter data before and after a 30‐min physical therapy program. Stride length, step length, cadence, walking velocity, single support time and double support time were measured. The range of motion of the hip, knee and ankle joint, and maximal trunk displacement on the horizontal plane were measured. Results:  In the high‐score group, significant improvement was observed in walking velocity, stride length and step length, and in the range of motion of the hip and knee joint. Maximal trunk displacement decreased significantly. In contrast, no significant improvement was observed in the low‐score group. Multivariate logistic regression analysis showed that Frontal Assessment Battery scores were a predictor of improvement in the range of motion of bilateral hip and knee joints, and maximal trunk displacement. Conclusions:  We showed that the subtests of motor learning of the Frontal Assessment Battery might be associated with the immediate effects of physical therapy on gait disturbance in Parkinson's disease. Geriatr Gerontol Int 2013; 13: 630–637.