Super-resolution in imaging through a transparent microlens has attracted lots of attention because of its exciting and promising experimental results with remarkable resolution improvement. However, the underlying theoretical foundation of the super-resolution still is not well understood. Here we use vectorial electromagnetic analysis to calculate the resolution through a microlens and find that traditional electromagnetic imaging theory does not lead to the experimentally verified high resolution record in literature. Therefore there must be some unknown mechanism behind the reported phenomena that hasn't been fully discovered.