In recent years there has been an increased demand for frequently updated Land Cover Classification (LCC) products. With ever greater numbers of Earth Observation (EO) and Remote Sensing (RS) platforms capturing data, the combined utilisation of data from multiple platforms has the potential to improve both the accuracy and frequency of LCC product updates through a process known as sensor fusion. This thesis examines how the fusion of RS sensors with diverse imaging characteristics can overcome challenges encountered when generating an annual LCC product. To undertake this examination, the technical specifications for the Second Generation Corine Land Cover (CLC+) Backbone raster product and its application on the island of Ireland is used as a case study for evaluating the use of mono-platform and sensor-fused RS datasets for LCC. A review of Machine Learning (ML) techniques in this thesis highlighted key factors crucial to achieving high-accuracy classification along with the proposal of a high-accuracy rapid inference Light Fully Convolutional Neural Network (LFCNN) architecture. This review also highlighted challenges when performing LCC on the island of Ireland, such as frequent cloud cover that can reduce the availability of optical satellite data, preventing the widespread use of high-accuracy multi-temporal ML classification. In this thesis, the application of a decision-level fusion approach is demonstrated as a means of mitigating the issue of frequent cloud cover and ensuring full classification coverage while also increasing classification accuracy. The versatility of a decision-level fusion approach was further demonstrated through the fusion of aerial and satellite RS data to improve LCC in complex non-homogeneous regions such as urban environments. The findings of this research have direct implications not only for performing LCC on the island of Ireland but throughout Europe and beyond, with clear recommendations provided for the generation of LCC products using EO and RS data.