Abstract Slope channel systems, well-known for their significance in hydrocarbon exploration and sediment transport and deposition in deep-water settings, are inherently complicated in the rock record. To unravel the complexity, a better understanding of their internal organisation is needed. Here an integrated qualitative and quantitative examination of internal architecture and vertical facies successions is reported from the San Fernando slope channel system of the Rosario Formation (Upper Maastrichtian), Baja California, Mexico. The San Fernando system (ca. 7–9 km wide, 400 m thick) comprises four channel complex sets (CCSs), of which this study focuses on the best exposed-CCS-B. Geologic mapping and photomosaic interpretation indicate that CCS-B features an erosion-surface-bounded and coarse-grained lower succession that records braid-like channel fills at the channel belt axis. The overlying upper succession is characterised by fine-grained overbank or passive channel fills intercalated with conglomeratic sinuous channel fills at the channel belt axis. By contrast, fine-grained overbank deposits and a capping hemipelagic interval (abandonment) dominate the upper succession at the channel belt margin. Visual inspection of the measured sections from CCS-B suggests multiple fining-upward packages, which are partly confirmed by stochastic analysis of vertical facies transitions using the Markov chain technique. This technique also quantitatively demonstrates that distinct preferred vertical facies transitions characterise different stratigraphic levels (lower succession vs. upper succession), and varying depositional environments (axis vs. margin) within the channel belt of CCS-B. The results of this study merit close consideration as analogues to submarine channel systems characterised by a lower braided pattern and/or an upper meandering pattern, and should find wide applicability in subsurface prediction of reservoirs, in other outcrop-based studies, and in constraining transition probability-based stochastic geological modeling.