This paper presents a new multi-channel integration method, called Single-Diagram-Enhanced multi-channel integration, and its implementation in the multi-purpose event generator MadEvent, which is based on MadGraph. The method allows MadGraph to automatically identify all relevant subprocesses for a given process, generate the necessary amplitudes and phase-space mappings, and pass them to MadEvent. MadEvent then produces a stand-alone code that enables users to calculate cross sections and generate unweighted events in a standard output format. The method is particularly useful for processes with many parton subprocesses, such as vector boson production in association with jets, and for handling complex phase-space integrations. The key idea of the method is to decompose the squared amplitude into a sum of individual Feynman diagram contributions, each of which can be efficiently integrated using a single channel. This approach reduces the complexity of the integration and allows for efficient computation by leveraging the structure of the Feynman diagrams. The method also enables the use of parallel computing, as each channel can be integrated independently. Additionally, the method allows for reweighting of channels based on their contribution to the final result, improving the efficiency of the integration. The paper provides several examples of processes relevant to current and future colliders, including vector boson production in association with jets, heavy-quark pair production, and Higgs production in association with top-quark pairs or W bosons. The results are compared with those from other event generators, showing good agreement. The method is shown to be effective in handling complex processes with many particles in the final state, and the package is currently limited to a certain number of Feynman diagrams per subprocess. The authors also discuss future improvements, including the factorization of amplitudes to reduce computational complexity and the extension of the method to handle a larger number of QCD partons. The paper concludes that the new method provides a powerful tool for the automated generation of event samples and the calculation of cross sections in high-energy physics.This paper presents a new multi-channel integration method, called Single-Diagram-Enhanced multi-channel integration, and its implementation in the multi-purpose event generator MadEvent, which is based on MadGraph. The method allows MadGraph to automatically identify all relevant subprocesses for a given process, generate the necessary amplitudes and phase-space mappings, and pass them to MadEvent. MadEvent then produces a stand-alone code that enables users to calculate cross sections and generate unweighted events in a standard output format. The method is particularly useful for processes with many parton subprocesses, such as vector boson production in association with jets, and for handling complex phase-space integrations. The key idea of the method is to decompose the squared amplitude into a sum of individual Feynman diagram contributions, each of which can be efficiently integrated using a single channel. This approach reduces the complexity of the integration and allows for efficient computation by leveraging the structure of the Feynman diagrams. The method also enables the use of parallel computing, as each channel can be integrated independently. Additionally, the method allows for reweighting of channels based on their contribution to the final result, improving the efficiency of the integration. The paper provides several examples of processes relevant to current and future colliders, including vector boson production in association with jets, heavy-quark pair production, and Higgs production in association with top-quark pairs or W bosons. The results are compared with those from other event generators, showing good agreement. The method is shown to be effective in handling complex processes with many particles in the final state, and the package is currently limited to a certain number of Feynman diagrams per subprocess. The authors also discuss future improvements, including the factorization of amplitudes to reduce computational complexity and the extension of the method to handle a larger number of QCD partons. The paper concludes that the new method provides a powerful tool for the automated generation of event samples and the calculation of cross sections in high-energy physics.