The low image quality of Electrical Impedance Tomography (EIT) hinders it from performing quantitative analysis in various fields like tissue engineering. This paper presents a gradient-guided multi-modal image reconstruction algorithm (named GGIR) for high-quality EIT image reconstruction. In the GGIR, a carefully designed first-order gradient-based regularization term serves as the interface incorporating structural information from the auxiliary image. The Laplacian regularization is also included in the GGIR for image smoothing. The performance of the proposed GGIR is evaluated by numerical simulation and real-world experiments. Compared with given single-modal and multi-modal algorithms, the proposed GGIR is superior in image quality improvement with a 40.2% ∼ 90.3% reduction in Err and a 16.5% ∼ 180.2% increase in MSSIM in the numerical simulation. Visual results in the simulation study and real experiments also prove the superiorty of the proposed GGIR. These evidences indicate that the GGIR has the potential to be applied to complex tissue engineering applications and regenerative medicine.