A single-bit spatial spectrum estimation method based on logistic regression relates to a spatial spectrum estimation field in array signal processing and a logistic regression field in artificial intelligence. The single-bit spatial spectrum estimation method settles problems of high calculation amount and relatively low precision in a traditional spatial spectrum algorithm on the condition of single-bit extreme quantification and ultralarge-scale antenna array. According to the method of the invention, single-bit received data are modeled for obtaining a sample model, and furthermore an observation model is converted to a real number domain for facilitating subsequent processing. After modeling, a spatial spectrum is used as a coefficient of a linear classifier; a flow pattern is used as an input sample; an array observation output is used as the output which corresponds with the input sample; and spatial spectrum estimation is converted to a linear classification problem. According to the algorithm of the invention, the linear classification problem is solved by means of a logistic regression, thereby obtaining a classification coefficient and a spatial spectrum which is generated correspondingly with an array input signal. The single-bit spatial spectrum estimation method is mainly used for spatial spectrum estimation.