Reducing the number of data recording points in a seismic data acquisition can reduce the cost of seismic operation. Nevertheless, this can reduce the quality of seismic data and may cause spatial aliasing. Jitter sampling data acquisition method can address this problem. In this method, the seismic data are recorded randomly and the missing data are reconstructed using some data reconstruction techniques. Conventional jitter sampling methods assume source and receiver locations are on a regular grid. In reality, this assumption is not valid due to the surface obstacles. In this paper, a novel method based on non-uniform Fourier transform is proposed to solve this problem. With this method, seismic data can be recorded anywhere with any sampling distance on a seismic data acquisition survey.