The existing modulation recognition algorithms for a pulse compression radar (PCR) signal can hardly adapt to complex modulation types and low signal-to-noise ratio (SNR). To solve the problems, with respect to the seven kinds of widely used PCR signals—including linear frequency modulation signal, Baker code, Frank code, P1 code, P2 code, P3 code, and P4 code—a modulation type recognition algorithm based on integrated quadratic phase function (IQPF) and fractional Fourier transform (FrFT) is proposed. First, signals are preclassified according to their chirp rates (CRs) estimated through IQPF. Then, FrFT is carried out depending on the order, which is correlated to the estimated CR. Finally, signals in each class are subdivided and modulation recognition is accomplished according to the features of the FrFT spectrum. The simulation results validate the feasibility of the algorithm. They also demonstrate that, compared against existing research, the proposal achieves better correct recognition performance for various modulation types under low SNR condition.