Despite increasing interest in generalized anxiety disorder (GAD) it remains one of the least studied anxiety disorders and little is known about its neurobiology. The aim of the current research study was to investigate the neural basis of GAD from two perspectives: fear generalization and anticipation of affective outcomes. Thirty-two women with GAD (17 with and 15 without comorbid major depression disorder) and 25 healthy women underwent functional magnetic resonance imaging (fMRI) while completing two tasks: a fear generalization task in which participants were presented with a conditioned stimulus (CS) that co-terminates with a brief electric shock, and generalization stimuli (GS) that range in perceptual similarity to the CS, and an anticipation task in which predetermined cues warned participants of upcoming aversive, positive, `uncertain' (either aversive or positive) or neutral movie clips. For the generalization task, neural reactivity in the insula, anterior cingulate cortex, supplementary motor area and caudate followed a generalization gradient, with a peak response to the CS that declines with greater perceptual dissimilarity of the GS to the CS; consistent with participants' self-report and autonomic responses for each stimulus. In contrast, reactivity in the ventromedial prefrontal cortex (vmPFC), a region associated with fear inhibition, showed an opposite response pattern (i.e., largest response to the GS most dissimilar to the CS). Patients with GAD, compared to healthy individuals, exhibited a flatter vmPFC gradient suggestive of deficient recruitment of vmPFC during fear inhibition. Anticipation of all affective clips engaged a common set of regions - the insula, dorsal anterior cingulate cortex, thalamus, caudate, prefrontal cortex and inferior parietal area - involved in various preparatory processes such as sustained attention, increased arousal, appraisal, and regulation. The nucleus accumbens (NAcc) and medial prefrontal cortex (mPFC), regions implicated in reward processing, were selectively engaged during anticipation of positive clips and the mid-insula was selectively engaged during aversive anticipation. Anticipation during the `uncertain' condition reflected a preparatory response for both aversive and positive stimuli showing activation in the NAcc, mPFC and mid-insula. Patients with GAD, compared to healthy individuals, exhibited enhanced reactivity during anticipation of neutral clips and a selective increase in visual cortical activation during affective anticipation consistent with a less discriminant anticipatory response. Findings underscore two processes - deficits in fear inhibition of stimuli that resemble conditioned danger cues and excessive anticipation of innocuous stimuli - which may contribute to symptoms of worry and anxiety in GAD.