We demonstrate and theoretically analyze the two-dimensional spatiotemporal focusing of femtosecond pulses by utilizing a two-dimensional spectral disperser. Compared with spatiotemporal focusing with a diffraction grating, it can achieve widefield illumination with better sectioning ability for a multiphoton excitation process. By utilizing paraxial approximation, our analytical method improves the axial confinement ability and identifies that the free spectra range (FSR) of the two-dimensional spectral disperser affects the out-of-focus multiphoton excitation intensity due to the temporal self-imaging effect. Based on our numerical simulation, a FSR of 50 GHz is necessary to reduce the out-of-focus two-photon excitation by 2 orders of magnitude compared with that in a grating-based spatiotemporal focusing scheme for a 90-fs excitation laser pulse. We build a two-dimensional spatiotemporal focusing microscope using a virtually imaged phased array and achieve an axial resolution of 1.3 μm, which outperforms the resolution of conventional spatiotemporal focusing using a grating by a factor of 1.7, and demonstrate better image contrast inside a tissue-like phantom.