To investigate the blind estimation of the symbol rate of a long code direct sequence spread spectrum (DSSS) signal, an algorithm based on autocorrelation processing is proposed. First, the second order moment of the autocorrelation estimator is used to analyze the long code DSSS signal. Next, the estimator output is further processed to get rid of the peaks at multiples of the pseudo-noise (PN) sequence period. Finally, another autocorrelation estimator and the Fourier transform are used to extract the symbol rate. Theoretical analysis and simulation results show the efficiency of the proposed algorithm.