The Kessler Foundation Neglect Assessment Process (KF-NAP) is an assessment tool for unilateral spatial neglect (USN), which is the scoring method for the Catherine Bergego Scale (CBS) based on detailed instructions. This study is aimed at determining the reliability and validity of the Japanese version of the KF-NAP (KF-NAP-J), evaluating the improvement of neglect assessment with KF-NAP-J, and comparing it with the original CBS for subacute stroke patients. We assessed subacute stroke patients admitted to our intensive rehabilitation hospital. Two KF-NAP-trained occupational therapists (OTs) assessed 22 patients. Before implementing the KF-NAP at the hospital, two other OTs assessed the other 23 patients using the CBS. We evaluated the interrater reliability of the KF-NAP and CBS using intraclass correlation coefficients (ICC) for the total scores, weighted kappa statistics for each subscale, and internal consistency using Cronbach’s alpha. We assessed the validity of the KF-NAP against the Behavioral Inattention Test (BIT) and Functional Independence Measure (FIM) using Spearman’s correlation coefficient. The reliability of both the KF-NAP and CBS was excellent. The weighted kappa results demonstrated that each subscale was in better agreement with the KF-NAP than with the CBS. In the KF-NAP, all eight subscales in which weighted kappa could be calculated were in significant agreement, and two were almost in perfect agreement. The KF-NAP moderately correlated with the subscales of BIT and FIM representing USN and activities of daily living. The USN detection rates of KF-NAP and BIT in the KF-NAP group were 63.6% and 22.7%, respectively. These results suggest that the KF-NAP, as well as the CBS, is useful to assess USN, which strongly impacts the rehabilitation outcomes in subacute stroke patients.