Missing data frequently occur in many applied domains and pose serious problems such as loss of efficiency and unreliable results for various approaches. Many real applications require complete data, thus, the filling procedure is a mandatory and precursory pre-processing step. DTWBI is a previously proposed method to estimate missing data in univariate time series with recurrent data. This paper introduces an extension of DTWBI, namely eDTWBI. Firstly, we simultaneously find the two most similar windows to the sub-sequences before and after a gap using DTWBI. Secondly, we impute the gap by average values of the following and previous sub-sequence of the most similar values. Experimental results on three datasets show that our approach outperforms than seven related methods in case of time series having effective information.