Recognizing faces in low resolution and blurred images is common yet challenging task. Local Frequency Descriptor (LFD) has been proved to be effective for this problem and is extracted from a spatial neighborhood of each pixel of a frequency plane regardless of correlations between frequencies. To explore the frequency correlations and preserve low resolution and blur insensitive simultaneously, we propose Enhanced LFD (ELFD) in which information in space and frequency is jointly utilized so as to be more descriptive and discriminative than LFD. The selection of window size of short-term of Fourier transform adaptive to the testing image is also analyzed. In addition, linear weighting fusion of recognition results given by magnitude and phase is proposed. The experiments conducted on Yale and FERET databases demonstrate that promising results have been achieved by the proposed ELFD, adaptive window size selection and fusion scheme.