In dynamic light scattering measurements, noise often makes inversion of the autocorrelation function to obtain the particle size distribution unreliable. To obtain accurate particle size distributions from noisy dynamic light scattering data, a modified inversion method based on the original Tikhonov regularization algorithm is proposed. In the method, the noise in the data is considered an independent variable. During the inversion process the number of rows and columns of the coefficient matrix equation is increased to accommodate this. Finally, using the dimensions of the coefficient matrix, the poor particle size distribution data is separated from the recovered particle size distributions, reducing the influence of noise in the data. The particle size distributions recovered from the dynamic light scatteringdata show that the modified Tikhonov regularization inversion algorithm can give rise to improved accuracy compared with the original inversion algorithm, especially for low signal-to-noise ratio data.