Digitalis lanata cell cultures grown as small undifferentiated aggregates in suspension culture can be redifferentiated into green embryos that produce cardenolides. The possibility of using a statistical (Box–Wilson) experimental design to study the effects of four different variables on growth, differentiation, and cardenolide production of D. lanata tissue cultures are investigated. The results of the analyses were processed by linear regression analysis. Mathematical models explaining the effects of the variables were developed. The concentration of maltose and the NO3−-NH4+ ratio were found to be significant variables for both growth and cardenolide production. The size of the inoculum was also important.