This study examined associations between drinking and smoking prior to treatment (biochemically measured at baseline), alcohol and tobacco craving, and biochemical alcohol and tobacco use during the analog trial period. We conducted a secondary data analysis of a randomized clinical analog trial where participants with a Diagnostic and Statistical Manual, Fourth Edition Text Revision (DSM-IV-TR) diagnosis of alcohol dependence, abuse or reported heavy drinking, with a co-occurring DSM-IV-TR diagnosis of nicotine dependence, abuse or reported heavy use, who were not seeking treatment were recruited. A generalized estimation equation model for longitudinal binary outcomes was created (N = 34) to determine the predictive effects of baseline tobacco use, alcohol craving, and tobacco craving on alcohol use over the 4 weeks of the trial. Baseline smoking was significantly (*p < 0.05) associated with drinking over time [odds ratio (OR) = 3.09*], while baseline drinking was associated with smoking (OR = 4.17*). Baseline alcohol and tobacco craving were positively associated with smoking over time (OR = 3.21* and OR = 1.92*, respectively) but were negatively associated with alcohol use over time (OR = 0.79* and OR = 0.57*, respectively). Heavier use of either tobacco or alcohol preceding treatment may require more intensive interventions in order to reduce tobacco and alcohol use. Future trials designed to address mechanisms of behavior change in the context of novel treatments could promote a better understanding of the cross-rewarding effects related to the co-use of these substances and lead to the development of more integrated and appropriately intense treatments for individuals with concomitant tobacco and alcohol use disorders.