This paper presents a comparative analysis of widely adopted Model-Based Systems Engineering (MBSE) methodologies, with a particular focus on how they manage traceability and integrate preliminary design aspects. While MBSE offers a structured approach to system development through models, many methodologies fail to ensure consistency across model views or to adequately bridge the gap between high-level system modeling and early design phases such as sizing. The study evaluates ten methodologies using a set of qualitative criteria, including development phases, traceability implementation, design-level support, tooling, language, metamodel, and ontology usage. The analysis reveals recurring gaps, such as limited integration of design calculations, inconsistent traceability mechanisms, and insufficient semantic grounding. Based on these findings, a roadmap is proposed to enhance future MBSE practices by integrating sizing views, defining semantic frameworks, and strengthening traceability verification.