Abstract: Analog-to-information convertor (AIC) is a device which converts a sparse analog signal to compressed information. Down-sampled random demodulator is a major structure of AIC, which, however, exits the problems of low sampling recovery precision and low compression ratio. In order to improve compressed ratio and sampling recovery precision, an implementation for AIC via multiphase random sampling and sub-sampled FFT is proposed. It consists of multi-phase frequency divider, pseudo random number generator, analog-to-digital converter (ADC) and accumulator. This AIC uses the property that down sampling after circular convolution in frequency domain is equivalent to multiphase random sampling in time domain and then applying sub-sampled FFT, making it an effective way of signal compressing sampling. The structure of this AIC is simple and easy to implement. Experiment results show the effectiveness of the proposed method.