Analyzing the pore structure and sealing efficiency of carbonate cap rocks is essential to assess their ability to retain hydrocarbons in reservoirs and minimize leaking risks. In this contribution, the impact of diagenesis on the cap rocks’ sealing capacity is studied in terms of their pore structure by analyzing rock samples from Ordovician carbonate reservoirs (Tarim Basin). Four lithology types are recognized: highly compacted, peloidal packstone–grainstone; highly cemented, intraclastic-oolitic-bioclastic grainstone; peloidal dolomitic limestone; and incipiently dolomitized, peloidal packstone–grainstone. The pore types of cap rocks include microfractures, intercrystalline pores, intergranular pores, and dissolution vugs. The pore structure of these cap rocks was heterogeneously modified by six diagenetic processes, including calcite cementation, dissolution, mechanical and chemical compaction, dolomitization, and calcitization (dedolomitization). Three situations affect the rocks’ sealing capacity: (1) grainstone cap rocks present high sealing capacity in cases where compaction preceded cementation; (2) residual microfractures connecting adjacent pores result in low sealing capacity; and (3) increasing grain size in grainstones results in a larger proportion of intergranular pores being cemented. Four classes of cap rocks have been defined according to the lithology, pore structures, diagenetic alterations, and sealing performance. Class I cap rocks present the best sealing capacity because they underwent intense mechanical compaction, abundant chemical compaction, and calcite cementation, which contributed to the heterogeneous pore structures with poor pore connectivity. A four-stage, conceptual model of pore evolution of cap rocks is presented to reveal how the diagenetic evolution of cap rocks determines the heterogeneity of their sealing capacity in carbonate reservoirs.