Smallholder farmers theoretically produce and market their output to meet livelihood needs of the family under given set of resource constraints. In spite of numerous policy efforts to ensure that smallholder farmers in Ghana and rural districts in particular, exploit the economic potential of cassava production, many of them continue to produce for subsistence with limited participation in the markets. To assess and explain the levels at which smallholder cassava farmers participate in markets, the study applied the partial proportional odds model in the framework of Vector Generalized Additive model (VGAM), which comprehensively addresses conceptual, modelling and interpretational complexities implied by market participation (MP) theory. Data for the study were collected from 365 cassava farmers in Sekyere Central (SC) district of Ashanti region, Ghana. The results showed that only 32% of the farmers participated in the markets as Net sellers, while 45% and 23% participated as Autarkic producers and Net buyers respectively. The econometrics analysis revealed that household size, access to market information, age, education, market access, membership to famer or community organisation, off-farm income, farm income, livestock holdings, cassava output and cassava being cultivated as major crop, appeared statistically significant in explaining MP decisions. Five variables, namely income, livestock, produce, access to market information and cassava as major produce, failed the proportional odds assumption and entered the model unconstrained, producing regime specific estimates. These results are indicative of how conditional, on a given regime, farmers respond to changes in the factors that affect their MP decisions, in terms of intensity and direction. Farm level policies, that target cassava farm households’ participation in markets should consider regime specific strategies.