Crystalline silicon solar cells face inherent performance bottlenecks. Limited by single-bandgap optical absorption and inevitable non-radiative carrier recombination, their maximum photovoltaic efficiency is strictly constrained by the Shockley-Queisser thermodynamic limit. Two-dimensional van der Waals (vdW) heterostructures have emerged as promising optoelectronic materials. They own atomically smooth and dangling-bond-free interfaces, coupled with excellent broadband optical absorption, suppressed carrier recombination, high carrier mobility and electric-field-tunable band structures. These unique advantages make layered vdW materials highly suitable for developing next-generation high-performance optoelectronic and photovoltaic devices. Rational design and assembly of two-dimensional vdW heterojunctions can effectively advance the exploration of high-efficiency photovoltaic functional materials. This work theoretically constructs an all-inorganic two-dimensional vdW heterostructure consisting of lead-free Cs₂SnCl₄ perovskite and monolayer MoTe₂. Different from traditional lead-containing organic-inorganic hybrid perovskites, the inorganic Cs-Sn based framework completely removes toxic lead elements. It also solves the structural instability caused by volatile organic cations, greatly improving the material reliability. First-principles calculations based on density functional theory are performed to systematically investigate the electronic structure and optical absorption properties of the Cs₂SnCl₄/MoTe₂ heterostructure. The simulated results verify that the designed heterostructure exhibits a typical type-II band alignment with an indirect bandgap of 1.26 eV. Benefiting from the interfacial optical synergy between the two functional layers, the heterojunction achieves outstanding optical absorption performance. Its optical absorption coefficient is higher than 10⁵ cm^-1 in the full visible spectral range. A slight decline of absorption coefficient is observed near the absorption edge, nevertheless. This weak optical attenuation, compared with pure monolayer MoTe₂, mainly stems from the intrinsic indirect bandgap feature of the composite heterostructure. Deformation potential theory calculations confirm that the Cs₂SnCl₄/MoTe₂ heterostructure possesses a favorable hole mobility of approximately 500 cm²·V^-1·s^-1, which ensures efficient carrier transport inside the layered structure. Even with excellent optical and transport properties, the pristine heterostructure still suffers from poor intrinsic photovoltaic performance. Large donor-acceptor band offset and obvious conduction band compensation easily trigger severe photogenerated carrier loss, which significantly reduces the intrinsic photoelectric conversion efficiency. External electric fields ranging from 0.1 V/Å to 0.2 V/Å are adopted to modulate the optoelectronic behaviors of the heterostructure. The applied vertical electric field can effectively adjust the band structure, gradually reduce the donor bandgap, and weaken the conduction band compensation effect of the type-II system. The heterostructure achieves a maximum photoelectric conversion efficiency of 4.13% at the optimal electric field of 0.11 V/Å. Once the applied electric field exceeds this critical value, the heterostructure undergoes a band transition from type-II to type-I configuration. The transformed band structure fails to realize effective spatial separation of photogenerated electrons and holes, which makes the heterostructure incapable of photovoltaic operation. In summary, the unmodified Cs₂SnCl₄/MoTe₂ heterostructure has low intrinsic conversion efficiency and cannot be directly applied to high-efficiency solar cells. Fortunately, external electric field modulation can effectively regulate band alignment, optimize the donor bandgap and alleviate conduction band compensation. This feasible band engineering method can prominently improve the photovoltaic performance of the heterostructure, providing a meaningful reference for the future design of high-efficiency two-dimensional solar cells.