with a as (
select '20160920' 날짜, 1 휴일구분 from dual union all
select '20160921' 날짜, 1 휴일구분 from dual union all
select '20160922' 날짜, 1 휴일구분 from dual union all
select '20160923' 날짜, 1 휴일구분 from dual union all
select '20160924' 날짜, 2 휴일구분 from dual union all
select '20160925' 날짜, 2 휴일구분 from dual union all
select '20160926' 날짜, 1 휴일구분 from dual union all
select '20160927' 날짜, 1 휴일구분 from dual union all
select '20160928' 날짜, 1 휴일구분 from dual union all
select '20160929' 날짜, 1 휴일구분 from dual union all
select '20160930' 날짜, 1 휴일구분 from dual union all
select '20160931' 날짜, 1 휴일구분 from dual )
SELECT 날짜 FROM (
select 날짜 , DENSE_RANK() OVER (ORDER BY 날짜) RN
from a
WHERE 휴일구분 = 1 AND 날짜 > '20160921' ) --to_char(sysdate, 'yyyymmdd')
WHERE RN = 5;

이러면 되지 않을까요?? 휴일 구분이 1 이고 입력한 날짜보다 큰 애들 중에 5등인 애를 뽑으면 될거 같단 생각을 해서요.. 

WITH TMP AS (
SELECT '20160920' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160921' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160922' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160923' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160924' AS DT, '2' AS GB FROM DUAL
UNION ALL SELECT '20160925' AS DT, '2' AS GB FROM DUAL
UNION ALL SELECT '20160926' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160927' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160928' AS DT, '1' AS GB FROM DUAL
UNION ALL SELECT '20160929' AS DT, '1' AS GB FROM DUAL
)
SELECT *
FROM (
    SELECT T.*
        , LEAD(DT, 5) OVER(ORDER BY DT) AS DT5
    FROM TMP T
    WHERE GB = '1'
) WHERE DT = '20160921'

Dev03 님 방법은 전체범위 처리가 되어 버리겠네요.

도미노 같은 놈 님의 방법이 부분범위 처리가 가능합니다.

Dense_Rank 보다는 Row_number 가 의미상 맞겠네요.

어차피 날짜가 유일하므로 결과는 같겠지만.