Seismic inversion is a fundamental procedure that converts seismic data into useful information about underlying rock and fluid characteristics. However, because seismic data are band-limited, the inversion process is intrinsically difficult, resulting in non-unique solutions. To overcome these issues, several constraints are used to enforce properties such as smoothness and sparsity on the inversion results. We propose a technique that includes the clustering properties of previous information, such as well logs and geological data, into the inversion process. This grouping helps to preserve geological continuity and improves the resolution of the inversion data. By incorporating this strategy into our inversion framework, we can better describe the subsurface and deliver more consistent findings. Our technique was evaluated on both synthetic and actual seismic data, confirming its ability to generate accurate acoustic impedance models. Furthermore, the approach generated deconvolved and denoised versions of the seismic data, which are useful for future interpretation. The membership sections generated by the inversion method also demonstrated considerable promise for tracing geological horizons, discriminating between distinct sequences and layers, and even predicting likely layer contents. In conclusion, this work proposes an upgraded seismic inversion approach that utilizes the ability of clustering to incorporate earlier geological knowledge, resulting in more accurate and interpretable findings.