The reconstruction of gapped microtremors recordings with the aid of CLEAN algorithm, which originally developed for spectral analysis of unevenly spaced time series, is studied. As proved in previous relative studies the algorithm has the ability to "clean" the spectrum of a discrete signal from artifacts and additionally (through an Inverse DFT) to reconstruct the signal in time domain. We studied the algorithm's effectiveness of reconstruction in microtremors recordings. We saw that for continuous gap with length up to 10% of the total length of microtremor recording, the algorithm was able to reconstruct efficiently.