d1 = pd.DataFrame({
   'DATE_AND_TIME':['2018-12-05 08:31:23.479', '2018-12-05 08:31:25.654',
                    '2018-12-05 08:31:27.013', '2018-12-05 08:31:30.004',
                    '2018-12-05 08:31:32.479', '2018-12-05 08:31:33.079',
                    '2018-12-05 08:31:34.504', '2018-12-05 08:31:37.729',
                    '2018-12-05 08:31:38.854', '2018-12-05 08:31:40.654'],
   'DELAY':[1, 2, 2, 3, 2, 2, 3, 2, 2, 2],
   'CUR':['EUR', 'EUR', 'EUR', 'USD', 'USD', 'EUR', 'EUR', 'USD', 'USD', 'EUR']
})
d1['DATE_AND_TIME'] = d1['DATE_AND_TIME'].astype('datetime64[ns]')
d2 = pd.DataFrame({
   'CUR':['USD', 'USD', 'USD', 'USD', 'EUR', 'EUR', 'EUR', 'EUR'],
   'DATE':['2018-12-04','2018-12-06','2018-12-09','2018-12-11',
           '2018-12-03','2018-12-07','2018-12-09','2018-12-11']
})
d2['DATE'] = d2['DATE'].astype('datetime64[ns]')

d1['PRE_ISD'] = d1['DATE_AND_TIME'] + pd.to_timedelta(d1.DELAY, unit='D')

CUR            DATE_AND_TIME  DELAY         ISD
0  EUR  2018-12-05 08:31:23.479      1  2018-12-06
1  EUR  2018-12-05 08:31:25.654      2  2018-12-08
2  EUR  2018-12-05 08:31:27.013      2  2018-12-08
3  USD  2018-12-05 08:31:30.004      3  2018-12-10
4  USD  2018-12-05 08:31:32.479      2  2018-12-08
5  EUR  2018-12-05 08:31:33.079      2  2018-12-08
6  EUR  2018-12-05 08:31:34.504      3  2018-12-10
7  USD  2018-12-05 08:31:37.729      2  2018-12-08
8  USD  2018-12-05 08:31:38.854      2  2018-12-08
9  EUR  2018-12-05 08:31:40.654      2  2018-12-08