Naturally fractured reservoirs occur worldwide, and they account for the bulk of global oil production. The most important impact of fractures is their influence on fluid flow. To maximize oil production, the characterization of a fractured reservoir at the scale of an oil field is very important. For fluid transport, the critical parameters are connectivity and transmittivity plus orientation. These can be related to fracture spacing, compliance, and orientation, which are the critical seismic parameters of rock physics models. We discovered a new seismic technique that can invert for the spatially dependent fracture orientation, spacing, and compliance, using surface seismic data. Unlike most seismic methods that rely on using singly scattered/diffracted waves whose signal-to-noise ratios are usually very low, we found that waves multiply scattered by fractures can be energetic. The direction information of the fracture multiply scattered waves contains fracture orientation and spacing information, and the amplitude of these waves gives the compliance. Our algorithm made use of the interference of two true-amplitude Gaussian beams emitted from surface source and receiver arrays that are extrapolated downward and focused on fractured reservoir targets. The double beam interference pattern provides information about the three fracture parameters. We performed a blind test on our methodology. A 3D model with two sets of orthogonal fractures was built, and a 3D staggered finite-difference method using the Schoenberg linear-slip boundary condition for fractures was used to generate the synthetic surface seismic data set. The test results showed that we were able to not only invert for the fracture orientation and spacing, but also the compliance field.