Brain metastasis (BM) is increasingly diagnosed in Her2 positive breast cancer (BC) patients. Lack of effective treatment to breast cancer brain metastases (BCBMs) is probably due to inability of the current therapeutic agents to cross the blood-brain barrier. The central nervous system (CNS) response rate in BCBM patients was reported to improve from 2.6%-6% (lapatinib) to 20%-65% (lapatinib in combination with capecitabine). Lapatinib is a poor brain penetrant. In this study, we evaluated the CNS penetration of capecitabine and hoped to interpret the mechanism of the improved CNS response from the pharmacokinetic (PK) perspective. Capecitabine does not have antiproliferative activity and 5-fluorouracil (5-FU) is the active metabolite. Capecitabine was orally administered to mouse returning an unbound brain-to-blood ratio (K p,uu,brain ) at 0.13 and cerebrospinal fluid (CSF)-to-unbound blood ratio (K p,uu,CSF ) at 0.29 for 5-FU. Neither free brain nor CSF concentration of 5-FU can achieve antiproliferative concentration for 50% of maximal inhibition of cell proliferation of 4.57 mM. BCBM mice were treated with capecitabine monotherapy or in combination with lapatinib. The K p,uu,brain value of 5-FU increased to 0.17 in the brain tumor in the presence of lapatinib, which is still far below unity. The calculated free concentration of 5-FU and lapatinib in the brain tumor did not reach the antiproliferative potency and neither treatment showed antitumor activity in the BCBM mice. The CNS penetration of 5-FU in human was predicted based on the penetration in preclinical brain tumor, CSF, and human PK and the predicted free CNS concentration was below the antiproliferative potency. These results suggest that CNS penetration of 5-FU and lapatinib are not desirable and development of a true CNS penetrable therapeutic agent will further improve the response rate for BCBM.